不合时尚的追求- Freeman J.Dyson (转载自水木)

发信人: happyzhe (good boy), 信区: AI
标  题: 转文：不时兴的人和不时兴的思想，常常对科学的进步有决定性的意义
发信站: 水木社区 (Sun Apr  5 16:50:47 2009), 站内



不合时尚的追求

Freeman J.Dyson

   一、引言 今天，我很高兴以高等研究所代表的身份，向Ｈumboldt基金会的会友讲话，因为高等研究所和Humboldt基金会都在努力支持国际范围的科学研究，又都面临着同样的困境和难处。我们正试图坚持150年前von Humboldt所开创的传统。为了对von Humboldt有所了解，我查阅了1910年版的不列颠百科全书，看到科学史家Ａgnes Clerke写的极漂亮的文章，如果你查以后的版本，则只能读到Clerke文章的片断。Clerke在文中描述了von Humboldt建立第一个国际气象和磁力观测网的工作，结尾铿锵有力："国际间的科学协作，乃是现代文明最富丽的硕果，而正是Humboldt的努力，成功地促成了第一次合作。"高等研究所和Humboldt基金会正以von Humboldt为榜样，尽力在我们自己的时代，加深和扩大国际间的科学协作。

    二、科学研究的时尚 我决定谈谈科学中的时尚问题，因为对于科学，特别是对于高等研究所和Humboldt基金会，这是个严肃而日趋重要的问题，我首先谈在高等研究所里看到的时尚；然后讲我们能从跨越漫长时期的科学史中吸取的教训；最后，就今后如何更明智地对待时尚说几句话。

    有种说法总是对的，而且在今天比以往任何时候都更加真切，即：对于能力一般的年轻科学家，最聪明的办法是追随占优势的时尚。任何一名青年科学家，要是没有杰出的才华，也没交上难得的好运，他首先关心的是找到一项工作并保有它。为此，他必须涉足于某个科学领域--它是控制着职业市场的、占据高位和有影响的权威们感兴趣的，并从事一项自己能胜任的工作。这些权威认为重要的科学问题，几乎就可以定为合时尚的问题。当然，给予工作的权力，在今天一般不由单个权威掌握，而由一个权威组成的委员会控制。但是，跟个人相比，委员会更难从一个时代的潮流中解脱出来。所以，关心自己生存的的青年科学家倾向于顺着踏就的路前进，这是毫不奇怪的。那些第一流的高级学术机构，向能轻车熟路跟随时尚的人提供保证，给予升职晋级，对不追随时尚者则只提供极少的机会。

    我们研究所也不例外，三十四年前，我首次来这里作访问成员。当时主要的权威是Robert Oppenheimer。他决定物理学中的哪些领域值得搞，他的口味总是跟当时最时兴的方向吻合。我那时年纪轻轻、雄心勃勃，拿着一篇讨论时髦问题的急就篇找到他，很快得到了一个永久的职位。这是那时的状况，今天也依旧如故。有些了解研究所历史的人可能反对上述看法，他们会说，研究所毕竟也给了Kurt Gödel 一个永久职位，情况确是如此。Gödel乃是本世纪少有的几个名不虚传的天才之一。在我们的同事中，他是唯一能跟Einstein以平等地位一起散步、聊天的人。Gödel从事非常深刻但不合时尚的一个数学领域的研究。随着年龄的增加，他显得更加赶不上潮流。我们研究所有理由为给他在教授会提供席位而骄傲。只有一个事实使这和光荣减色：Gödel自来研究所生活和工作，从普通成员升到教授竟花了十四年时间。Gödel有如此的独立和潮流精神，才使得我觉得，研究所在经历十四年的踌躇之后，终于使他成为一名教授，总算也是值得一提的一点功劳吧。晚做总比不做强！

   今天来研究所工作的青年物理学家，比起三十年前的我，受到更大的压力。 首先，他们多半是靠跟政府的合同得到钱的，合同约束他们在确定的时间内从事指定 科学领域的研究。当然，我们不必过份从字面上理解“合同”这个措词。 管理合同的国家科学基金会和能源部的官员都是理智的， 允许我们对所承担的义务作带点伸缩性的解释。 如果研究所内某些靠合同挣钱的成员，打算搞跟合同无关的课题， 那时没有人会强迫我们把这些人赶到大街上去。 有些人的兴趣所在不适于签入合同，他们的工作一般就由研究所基金会支持。 但不管怎么说，合同仍是严肃的，具有约束力， 它从总体上规定了研究所物理部的访问成员应积极从事的工作领域。 合同确定了物理学的主流应该是什么。我们对邀请来工作的成员， 必然要求他们的工作能容易地纳入这一项或那一项合同之中。

    三十年后的今天，我也成了权威中的一份子。我努力鼓励年轻的物理学家在非时尚的领域搞研究，但只能以一种既不道明又十分无力的方式进行。我试图让很少几个没有合同支持的研究领域保持生气。我力图让研究所的大门向具有独立思想和逆潮流的人物敞开。我要始终开着一扇门，以待另一个Kurt Gödel找上门来。不过我不得不承认，我企图阻挡时兴潮流的努力，其效果跟杰出的前辈Ｃanute王[注１]阻止大西洋潮汐的结果差不多。今天的年轻人，被一种比合同或权威更强的力所驱赶，追求着时兴的玩意儿。这股驾驭年轻人时髦的力量就是同辈的压力，就是追赶时髦本身的刺激。他们知道舞台在那儿，并想登台表演。他们知道只有短暂的时间来证明自己是个科学家。他们知道在配给他们使用的短时间内做出有价值成就的最好办法是随大流，尽快地在已成熟的领域摘取科学果实。

    年轻的科学家们力争尽快获得成功、力争尽快取得报偿， 这本身并非坏事。他们的努力集中在一些时兴的专门领域，也不一定有害。 毕竟，时尚问题之所以成为时尚，并非由于象某些时装设计师那样的灵机一动所致， 而是大部分科学家认为它们重要。有一条普遍法则，大多数人的判断总是有根据的。 时兴的领域常常都是些在其中获得了极其重要发现的领域。 年轻科学家拥向这些领域，以期作出轰动世间的发现，这是无可非议的。 确实，在时兴领域中许多人同时研究一些课题， 大大增加了研究所日常生活中乐趣和激情。 对于你在时兴领域开发宝藏进的每一次小的成功、每一回短暂的凯旋， 朋友们都会在饭桌上或讨论班里谈起它。如果失去对时兴问题的共同旨趣， 如果没有这种对新鲜消息和传闻的关心，我们研究所的生活将变得十分乏味。

    那么，我为什么还不满足呢？ 我为什么要为那些年轻人--他们正做着我自己在那种年龄时做的事--鸣不平呢？ 我之所以有牢骚，因为我认为我们的工作不应该百分之百都是合潮流的。 时尚的研究是有用的、重要的和激动人心的。 我们可以为年轻一代搞时髦课题并有所建树而骄傲。 出于我能理解并尊重的理由，我们将看到他们中的大多数人会永远乐于搞时尚课题。 我的意思只是说，必须为少数不搞时尚研究的人留下位置。 我们应该发现那些不适于纳入流行款式的少数人，并对他们加以鼓励。 在为研究所选人时，必须稍微偏向一点非正统和不从习俗的人。 如果连我们这里都不给搞非时尚科学家的从业者一席工作之地，那么还有谁会给呢？！

    三、以往的历史 由于存在许多非时尚的科学， 支持它们的主要困难之一是选择问题。非时尚的科学千姿百态，没有任何统一的结构。 上星期，我穿过批Princeton 大学的Forrestsl Cumpus (一处校园)时， 遇到两个研究生静静地坐在草地中间，起初， 我以为他们正在享受阳光和八月午后的静谧。 可是走近时，我看到他们正全神贯注地做什么精巧的操作，手一点不能颤动， 精神不得丝毫分散。当走到跟前时，我才弄清他们正忙着把一小块铅粘在蜜蜂背上。 我静静地在旁观看，等他们做完全部工作，便跟着来到他们的实验蜂箱， 箱上装备有照相机和录相机。这两个人正在更精确地做Karl von Frisch 的经典实验， 并进一步扩充实验内容：原实验用于研究蜜蜂用舞蹈传递信息的系统。 他们已发现，当蜜蜂发现蜜源离巢相当远时，它们的舞蹈更明显、更有力、了更精确。 不幸，大多数蜜蜂只在蜂巢附近找蜜，返巢时只是简单地、马马虎虎地跳一阵。这两个学生想观察高精度的舞态，便设计一套办法让蜂表演更明显的舞姿。当一只蜜蜂负重45毫克铅时，只要飞一小段距离它就以为飞了很长时间。蜜蜂以所费的气力来感知飞行的距离。所以，负重的蜜蜂每次采蜜回巢后，都跳出精细的舞蹈。

    上面说的是典型的非时兴科学的例子。事情发生的地点就在我们Princeton大学的门口，我并非提议高等研究所应支持某个昆虫学学派。但蜜蜂实验说明，一切这类非时兴研究的特点，使得支持它们变得困难。它们的规模很小，研究对象各式各样，风格特异，看起来缺乏严肃性。

    为了说清非时尚科学具有真正的和持久的重要性，我来谈我擅长的领域：数学物理。数学物理是这样的人从事的学科，他们力图用严格的方式和纯数学的方法，达到对物理现象的深刻理解。这门学科处于物理和数学的交界处。数学物理学家的目标不是对现象进行数量方面的计算，而是从质的方面去理解。他们提出定理，加以证明，但不依赖数学和计算机。他们的目标在于用数学的精确性，阐明物理理论赖以确立的概念的含义。

    数学物理有三个性质使它跟眼下的讨论有特殊的关系。 第一，它为更实际的物理领域提供基本思想和专门语汇，它从大的范围阐明事物的性质，因而很重要； 第二，它的进展缓慢，一个新概念从创生到能有效的使用， 基本上经历五十到一百年之久； 第三，它几乎总是非时兴的，因为它的周期比科学浪潮的周期大约慢车10倍。 由于它不时髦，所以在欧洲对它的关注与支持，总比在美国强得多。

    有一位伟大的数学物理家的工作，对今日的物理学仍然无比重要，我指的是Sophus Lie,他已去世八十年了。他的伟大工作完成于十九世纪七十至八十年代，但只是在刚过去的二十年间，才支配了研究粒子的物理学家的思想。 Lie第一个理解并清晰地陈述了群理论可作为物理原理的起点。他几乎靠单枪匹马构造了浩大而漂亮的连续群理论，并预见到有朝一日它将成为物理学的一个基础。一百年后的今天，每个按照破缺或无破缺对称性研究粒子分类的物理学家，都自觉或不自觉地使用Sophus Lie的语言。可是当Lie在世时，他的思想并不合时尚，几乎没几个数学家理解它，更不用说物理学家了。Felix Klein 是为数很少的能理解和支持他的大数学家之一。

    Lie属于这样一种人，他们似乎承受着不公平的厄运， 1870年普法战争爆发时，年轻的Lie正在法国漂泊。他是挪威人操着带普鲁士口音的法语。枫丹白露的爱国者认定他是普鲁士奸细，把他投入监狱，由于法国战败，形势一片混乱，当Lie的法国朋友最终找到关他的牢房并成功地使他获释时，他正静居囚笼，搞出了新的数学发现（Lie, 1877）。在世纪交替之际，Rouse Ball出版的数学史中，以悲怆的语调结束对Lie工作的评述（Rouse Ball, 1908）： “看来，Lie一直很失望，因他的工作价值没得到普遍承认，他为此而苦恼……。在他生命的最后十年，他常陷入沉思，想着他被过份忽视了的过去，使他心情不快。”

    另一位伟大的数学物理天才是Hermann Grassmann, 他在世时比Sophus Lie更不合时尚。 1844年在Stettin当预科学校的教员时，他发表了题为Die Lineale Ausdehnungslehre（扩张演算）的著作，首次引入了向量、向量空间和反交换代数的基本概念。它们在二十世纪的物理学中极其重要，但在十九世纪时却不然。在他生活的世纪，Grassmann一直在那所不知名的预科学校当教员，科学院的权威对他不闻不问。不过，他比Sophus Lie 有更强的适应性。他不是老想着得不到数学家们的承认这件事，而是开辟第二战场，去学习梵文。他把Rig-Veda（印度古经典四吠陀之一）译成了德文，因而有了不小的名气。也许，如果命运安排你成了不被承认的数学天才，为了健康起见，去当个预科学校的老师比当大学教授要好一些。

    为准备这次讲演，我到研究所图书馆查过资料， 我高兴地发现了一本1878年版的Ausdehnungslehre(维数理论)， 标题页上用铅笔写的Minkowski的名字 --他是Einstein的老师， 第一个理解相对论的数学家。1878年出的这本书中有Grassmann写的序言（仍是在Stettin写的），他兴奋的表达了如下希望：新版本将比三十四年前的头一版受到学术界更多注意。序言之后有一行脚注："Der Verfasser ist während des Druckes gesorben"(本书付印时作者已去世)。 只是到了十九世纪九十年代，Felix Klein--一位在为非时尚的事业战斗时总是毫不吝啬气力的人，才促成了对Grassmann的正式承认，并出版了他的全集（Grassmann, 1844, 1878, 1894）。

    数学物理在更近期的一个伟大发现，是Hermann Weyl于1918年提出的规范场的思想。这一思想仅过了五十年就在现代基本粒子物理学中获得了地位。量子色动力学是1981年粒子物理学家最时髦的理论，从概念上看，它就是Lie的群论代数和Weyl的规范的综合。Weyl提出规范场时的情况，跟Lie群和Grassmann 代数发现时遭遇完全不同。Weyl既有名气，工作也得到了承认。 他在1918年搞的正是物理学中最时兴的领域：新诞生的广义相对论。 他创立规范场是为了解决将重力和电磁力统一起来的时尚课题。 几个月内，他的规范场变成最时髦的玩意儿。 然后Weyl和其他人发现，规范场的提出并没达到预期的效果， 即他们在事实上并不适于Weyl原来创立它们的目标。 它们很快又不时兴了，甚至几乎被人忘却。 又经过五十年漫长岁月之后，规范场在一个完全不同的方向上--量子电动力学及其在近期引出的量子色动力学方面的推广， 清楚地显示出它的重要性。为规范恢复名誉的关键一步，是由我们Princeton的同事Frank Yang和他的学生Bob Mills于1954年迈出的， 那是在Hermann Weyl 去世前一年的事（Yang和Mills，1954）。 没有证据说明Weyl知道或注意到Yang和Mills使用从他脑袋里蹦出来的娃娃所从事的研究。

    规范场的故事充满了讽刺意味。一个时髦的思想，本想用来解决某个问题，但这个问题本身是短命的。经受长期的冷落之后，规范场最终以物理学里程碑的雄姿屹立于世。在漫长的数学物理发展史上，不乏这种反复的例证。Hamilton发明的四元数，曾被欢呼为解决十九世纪物理问题的灵丹妙药。可是在世纪转折之际，因无用而被弃置。到本世纪二十年代，它有以量子力学中自旋矩阵的形式恢复了青春，现在，它又光荣地跃入了夸克场理论。 Gauss发明的微分几何，起初只是他从事测地学和绘制地图等实际工作的副产品，经天才的Riemamn之手，它被改造成一个具有抽象一般性的新天地；五十年后，又作为Einstein重力理论的基础立于世人面前。这些历史有一个共同点，它们都经经历一个漫长的时期，从发生到结束通常超过单个人的生命期，而最终的结果完全无法预知，发明具有决定性意义的概念的人中，没有一个能对最终使用这种概念的物理领域有些微的感知。

    往事讲得不少了，我想，我已经给各位充分的历史见证以证明我的论点： 不时兴的人和不时兴的思想，常常对科学的进步有决定性的意义。 现在该讲讲现实和未来我没有理由期待今后科学思想的发展格式跟过去不同。 我们能够期待，在未来的岁月，非时兴的思想显示其重要性的机会跟过去一样频繁， 当然，这要经过漫长的孕育期，并在人们所不熟悉的领域崭露头角。 作为科学进步的卫士，我们面临着如何识别有前途但不合时尚的思想以及 如何支持它们的问题。

    四、魔怪和教训 首先，让我们环顾数学世界， 看看能否鉴别出在二十一世纪可能成为物理学基本构件的非时髦的思想。 要是走运，我们说不定能挑出未来杰作的侯选者。 当然，不能奢望在我们的有生之年，就弄清这种挑选是否正确。

    粗略地讲，非时尚的数学就是Bourbaki的权威们宣布为不属于数学的那部分数学，许多非常漂亮的数学发现属于这一范畴。据Bourbaki的观点，一种思想要称得上是数学，应该是一般的、抽象的、统一的，并跟数学的其余部分有清晰的逻辑关系。被排除在数学之外的是特殊的事实和具体的对象，它们的存在缺乏相应的理由，数学农称之为偶然或散在的事物。非时尚的数学主要跟具有意想不到的妙处的对象有关，如特殊函数、特殊的数域、异常的代数、散在有限群。我劝诸君到数学中这些尚未系统化、尚未形成学科的部分，去寻找物理学下一次革命的火种。它们具备奇异性和意外性的品质。它们不容易纳入漂亮的Bourbaki的逻辑结构。正是基于上述理由，我们应该珍爱它们，去开发它们。请记住两年前我们的所长Harry Woolf在一次讲演中的基调，那是他在研究所纪念Einstein诞辰白周年纪念会上引用的Francis Bacon的一句话： "没有奇特的奇异性，也就不存在于不同的美丽"（Woolf,1980）。

    我将简要地谈谈散在有限群（Conway, 1980）的特殊的奇异性。 散在有限群的历史始于法国数学家Emile Mathieu，他在1861年发现了第一个这种群， 1873年又发现第二个。 跟通常获得这类发现的情形一样，Mathieu并不知道自己发明了散在群， 事实上，他的文章的标题中没有"群"这个词（Mathieu, 1861, 1873）。但是，他清楚的知道已找到某种非常漂亮和重要的东西。 用几何语言讲，我们可以说他已经发现在12和24维空间中， 存在一种具有奇特对称性的结构，但在任何维数不是12或24的空间中， 不存在这种结构，他的工作发表了，但在其后的一百年里并不时兴。 正如被公认的数学家喜欢说的那样，这是珍奇的孤品，没有开辟任何前进的道路。

    大约七十五年后，Mathieu群在编码业务中表现出某种实用的重要性。每个Mathieu群都给一种特别有效的纠错码提供了基础。当然，Mathieu群在实际中的应用并未招致数学家的青眯，他们的口味让Bourbaki给限制住了。

    接着的二十年间，风云突变，各方面的数学家用各种方法发现了新的散在群组成的宏大的"动物园"。他们之中有的是按照Mathieu的思想找到的；另一些是通过研究一个非时尚的问题引出的：把24维的台球尽可能紧地装进24维欧几里得空间（Leech, 1967）；还有的是在大计算机上试算排列组合问题时创造的。

这些发现有一些共同点：具体性、经验性、实验性和偶然性。这跟Bourbaki的精神正好相违。包括Mathieu的结果在内，总共发现了25个散在群。与此同时，群论专家的团体，用更一般和抽象的方法，成功地证明了散在群的总数不能大于26，所以，两年前的形势是还剩下一个散在群可寻。当时知道，如果着最后一个群存在，它将是所有散在群中最大和最漂亮的（Conway和Norton, 1979）。正在猎取它的人给它起了绰号叫"魔怪"或"亲密的巨兽"。

    去年，当Bob Griess从Michigan大学来高等研究所访问时，上面的故事终于有了结尾，他找到了构造这个魔怪的方法（Griess, 1981）。正巧在昨天，我收到从Michigan几来的长篇论文的最后部分，其中包括对他的研究的完全和肯定的评价(Griess, 1982)。对于那些不辞辛劳地从细节上弄清Bob Griess构造的人来说，魔怪的面目已暴露无遗，他们可以感到满足和快意。现在，这最后的也是最大的散在群已无懈可击地独立于世，成为一座不朽的纪念碑。

    这一切对物理学有什么意义呢？也许，一无所有。也许，散在群只是数学史上一弯可爱的滞水，远离浩荡主流的奇妙的插曲。我们绝没发现一点儿迹象，说明物理世界中的对称以任何方式跟散在群的对称发生联系。迄今，我们所知道的是，不管有无散在群存在，物理世界的面貌和功能依然如故。但是，我们不应过分地肯定它们之间无关。缺乏证据跟不存在证据到底不是一回事。在物理学史上，有过比意想不到的散在群的出现更奇怪的事。我们应该永远准备好迎接意外的事情。我必须坦白地招认，我内心存着希望，没有任何事实和证据支持的希望：有朝一日，在二十一世纪，物理学家将与魔群邂逅，以某种出人意料的方式将其纳入物理世界的结构。这只是一种莽撞的推测，几乎肯定是错的。有利于这一推断的唯一证据来自神学。这个强有力的证据是：宇宙的创造者喜爱对称。如果他喜爱对称，那么还有什么别的对称?quot;魔怪"的对称更可爱呢？

    散在群只是不合潮流的数学家创造的怪异而奇妙的思想宝库中的一例。我还能举出许多例证。你能想象一个正面体，--由完美的对称元构成的物体，--排列成完美的对称结构，总共有11个面吗？去年，我的朋友Donald Coxeter（在Toronto）找到了这个多面体（Coxeter, 1981）。 有朝一日，会不会发现Zeta函数的零点（Riemann在120年前猜测它们具有某些性质， 现今仍是数学中重要的秘密之一）跟物理世界有隐秘的联系呢？ 去年，Andrew Odlyzko(Bell实验室使用Cray计算机的数学家) 发现了Zeta函数零点的某些新的和出人意料的性质。 Kurt Gödel的不完全性定理（证明纯数学中存在这类问题，任意给定一组有限个公理和推理规则，都无法解答它），是否有一天会使我们对物理知识的限度有更深入的理解？不管你在哪个思想的王国游历，总会发现各种奥秘的暗示，听到有关藏匿着的各种事物间联系的传闻。
时间不多了。 我必须践约讲讲对支持科学研究的具体意见。我是针对高等研究所和Humboldt基金会讲的。这既是我们的义务，也是我们这两个比政府更具慧眼的独立组织的殊荣。我们应能采取一种比政治家和博士后的学生看的更远的科学观。目光远大的科学观教我们怎么做呢？从上面讲过的许多故事应引出什么教训呢？教训只有一条，很简单：应该更多地注意、更有力地支持非时尚的研究。在科学史上任一特定时期，最重要和最重要和最富成果的思想往往潜伏着不被利用，原因仅仅是它不合时尚。具体到数学物理领域，从新思想的孕育到它成为科学思想的主流，通常要磨蹭五十或一百年。如果这是衡量基础性进展的尺度， 那么结论必然是：在数学物理领域从事基本研究的任何人几乎肯定是不合时尚的。

    当然，我们不应该停止支持使大多数年轻科学家忙碌和高兴的时尚研究。 但我们应拨出一部分经费，也许是十分之一或四分之一， 以支持从事非时尚工作的不合潮流的人。我们不应该害怕看到做傻事， 或是看到一堆破烂；我们不应该害怕支持可能完全失败的冒险事业。 因为我们是独立的、我们有权利冒险和犯错误。 那些仅仅支持搞无危险、无犯错误机会的研究的机构，实际上只是支持了平凡的人。 如果我们靠良知和勇气，支持不时兴的人，去做正统观念认为是不对题和冒险的事， 这就提供一种好的机会，为科学拯救很难得到的Sophus Lie或是Hermann Grassmann。 当我们时代的所有时髦动人的成果早被人遗忘之后，他们的思想仍将驰名于世。

参考文献


Conway, J. H. (1980) Monsters and Moonshine, Then Mathematical Intelligencer, 2, No. 4, 165-171.
Conway, J. H. and S. P. Norton (1979) Monstrous Moonshine, Bull. London Math. Soc. 11, 308-339.
Coxeter, H. S. M. (1981) "A Symmetric Arrangement of Eleven Hemi-Icosahedron", to be published.
Grassmann, H. (1844, 1878, 1894) Die Lineale Ausdehnugslehre, 1st ed. (ott Wigand, Leipzig) 1844, 2nd ed. (Otto Wigand, Leipzig) 1878, 3rd ed. in Grassmann's collected works edited by F. Engel (Teubner, Leipzig) 1894.
Griess, R. L. (1981) A Construction of F_1 as Automorphisms of a 196883-dimensional Algebra, Proc. Nat. Acad. Sci USA, 78, 689-691.
Griess, R. L. (1982) The Friendly Giant, Invent. Math. 69, 1-102.
Leech, J. (1967) Notes on Sphere Packings, Can. J. Math. 19, 251-267.
Lie, S. (1877) letter to A. Meyer, published in Sophus Lie, Gesammelte Abhandlunger, ed. F. Engel (Leipzig, Teubner, 1922), Vol. 3, Anmerkungern, p. 691.
Mathieu E. L. (861, 1873) Mémoire sur l' éstude des functions de plusieurs quantities, J. de Math. Pures et Appliquées, 6, 241-323, "Sur la foncion cinq fois transitive de 24 qunatités", J. de Math. Pures et Appliquées, 18, 25-46.
Rouse, Ball, W. W. (108) A Short Account of the History of Mathematics, 4th ed. (MacMillan, London), p. 478.
Woolf, H., ed. (1980) Some Strangeness in the Proportion: A centennial Symposium to Celebrate the Achievments of Albert Einstein (Addison-Wesley, Reading, Mass.)
Yang, C. N. and R. L. Mills (1954) Conservation of Isotopic Spin and Isotopic Gauge Invariance, Phys. Rev. 96, 191-195.


译注1： Canute王：英格兰及丹麦的王（995-1035）。

原题：Unfashionable Pursuits。译自The Mathematical Intelligencer 5:3 (1983), 此报告是1981年８月２４日在Princeton高等研究所做的。
中文翻译：袁向东译，吴允增较。原载于《数学译林》。

The Business of Mining the Twitter Stream

The Business of Mining the Twitter Stream

(http://datamining.typepad.com/data_mining/2009/02/the-business-of-mining-the-twitter-stream.html)

February 19, 2009

While mining Twitter data for business and marketing intelligence (trend/buzz analysis, sentiment/opinion mining, authority/influence analysis) looks like a compelling path to explore for a business model, it is important to consider the proposition from the point of view of the customer. Enterprises have been working with vendors in this space (mining social media content for BI) for well over 5 years and already have expectations regarding the features and quality of reports that these analytics needs to deliver to be useful (actionable).

• Domain coverage: how broad is the topical space available in the solution? Crawling all data sources is the way to win here.
• Demographic coverage: the broader the demographic coverage (and the accuracy with which the demographic features of the content authors can be determined) the better.
• Content Analysis/Text Mining: how well does the solution take all the unstructured content and deliver structured interpretations that can then act as the input for further data mining. This is generally a matter of applied research (taking the current state of the art in text mining and making it work with the greater variety and complexity of social media content).
• Timeliness: how timely is the analysis. This is generally a function of how timely the data is collected. Blog data, for example, can be gathered in a very timely manner thanks to the ping/feed  mechanism. However, the reality of real time mining is that the consumer of the data is the real calibrator - real time may mean 4 hourly, not second by second.

If the business model for Twitter is going to be mining the Twitter stream for BI/MI, then they will be competing with companies that gather very large data sets (weblogs, usenet, message boards, reviews, groups, mailing lists, etc.). Seth Grimes suggested that the short texts of the Twitter stream may make hard problems like sentiment mining simpler as the limited space requires the author to be concise. However, this is a double edges sword as it means that the depth of analysis will be far shallower.

I believe that mining Twitter data will be a very exciting experiment, but I think that if Twitter goes down this path, it will have to either provide analytics over the other data sets, or partner with an existing company (say Visible Technologies). In fact, such a partnership would take the burden of building out an analytics engine away from the small Twitter team allowing them to continue to focus on infrastructure and ensuring the flow of this valuable data stream.

Add to del.icio.us • Email this

February 19, 2009 in socialmedia | Permalink

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c994053ef011278fcc1a428a4

Listed below are links to weblogs that reference The Business of Mining the Twitter Stream:

Comments

Matthew, thanks for the mention. I'd venture that tweet mineability is also easier because short messages cover a single topic.

Short messages are easy to post so they folks can post more frequently. So maybe the more interesting thing to mine from twitter is message propagation. Then from propagation threads and connectedness patterns, one could infer influence networks and knowledge about the types & topics & forms of messages that travel farthest and fastest.

Posted by: Seth Grimes | February 19, 2009 at 05:12 PM

I don't get it, anyone can mine twitter for sentiment (using the search API)... why would twitter reinvent the wheel?

Posted by: Roddy | February 19, 2009 at 07:14 PM

Nice analysis, Matthew - I'd also say that as part of the way that people use Twitter is to share links to interesting content/conversations elsewhere, the need to be analysing the networks around the Twitter streams is very important indeed.

Posted by: Antony Mayfield | February 20, 2009 at 03:16 AM

There's a reality of the value of the raw data to the marketplace, which I'll get to in a minute. Regardless, short messages may very well be harder to search, not easier. Here's some reasons why:

* For indexing purposes, it's not only the corpus of the text that matters, it's the number of objects. So a search architecture has to take that into account. It's a non-trivial problem; especially with the kind of volumes involved here. Not to mention that servers are going to be thrashed with reading/writing if anything is meant to be done real time. (Perhaps less so for batch analysis of course.)

* Next we have the nature of the messages themselves. Due to the 140 character nature, there's an increase in odd acronyms even beyond the brb, lol, etc. Perhaps synonym dictionaries could be produced, but the variability here seems extreme just based on anecdotal experience.

* Regarding sentiment mining, that's difficult enough in larger text, but may be harder in small text. Not for raw sentiment where the phrases are obvious. But sentiment analysis lags with regards to humor and sarcasm, which may need more markers to divine actual meaning.

These are solvable problems. And in the latter case, it might not matter that terribly much if some stuff gets missed as general trends can still be spotted easily enough. Personally, I feel confident someone(s) will work this out to some reasonable degree of satisfaction.

Next, as to the dollars. I can tell you from experience the industry does not value the raw data terribly highly for specific social media data streams. The value added analysis? Yes. The actual data? Not so much. This is because it's easy enough for a variety of people to crawl blogs, forums and so forth. And several do, though in some cases there's really only a couple of providers feeding data to the 60+ reputation monitoring companies.

Unless Twitter made itself the sole availability for the full data stream, they wouldn't be able to command that great a price. I'm just guesstimating based on past experience with other data types here, but MAYBE 1M / month if they sold to every rep services company out there. (Who would in turn add analysis and re-sell for much more.) That's decent money, but it's not 'to the moon' money. I could be wrong here. People are valuing this stuff more highly. But to really capitalize on it, there's no way they could just let anyone suck down all they could eat off the stream. Which means less open. Which is fine. They're entitled to do so.

We'll see!

Posted by: Scott Germaise | February 20, 2009 at 02:06 PM

Nice post. I have included this blog into my rss subscriptions. Very nicely put on data mining using social media. I honestly have not thought about it in this much detail but it makes sense and could be used as a great competitor intelligence tool!

Posted by: video Promotion | February 20, 2009 at 07:39 PM

I'm still working on it but Twitter data sure is tasty.. Lots of goodies !

Thanks for the run down'

Mike

www.wannadevelop.com

Posted by: wannadevelop.com | February 24, 2009 at 05:30 AM

There is a lot of potential in analyzing Tweets : Segmentation of users, Sentiment Analysis to name a few. In my experience, the fact that tweets are maximum 140 characters makes things easier in catching emerging trends but also in Text analysis.

Combining Information Extraction and Ontologies (using IE to mark Text and using NLP to insert information to an Ontological Setting) is the way to go although it requires considerable effort.

Posted by: Themos Kalafatis | March 03, 2009 at 05:12 PM

 

Comments on Face Recognition (zz from newsmth)

发信人: cbir (Fighting for honor), 信区: AI
标  题: dodo：人脸识别方法个人见解(zz from prfans)
发信站: 水木社区 (Tue Jul 10 14:00:00 2007), 站内

http://prfans.com/forum/redirect.php?tid=97&goto=lastpost#lastpost

注：这个帖子很精彩，应该是prfans的震坛之宝:)，没有经过作者同意转载到这里，猜测转载到这里应该不违背作者初衷，呵呵。

dodo：人脸识别方法个人见解


看到j.liu关于人脸识别的帖子，萌发写这个帖子的念头。没有别的意思，就是想抛砖引玉，把问题说的全面一点，希望j.liu和回其帖子的兄弟姐妹们不要介意。如有兴趣，欢迎继续讨论。在以下讨论中，

TPAMI = IEEE Transactions on PAMI 这个杂志
PAMI  是指 pattern analysis and machine intelligence这两个领域


1）PCA和LDA及其相关方法

Eigenfaces和Fisherfaces无疑是人脸识别中里程碑式的工作。就使用的方法而言，PCA和LDA都不是新方法，但是他们都是被第一次十分明确的用在人脸识别中的方法。之所以说"十分明确"，是因为就发表的时间来看，这两个论文都不是首次把这两个方法用在PAMI相关的分类识别中。这给我们一个小小的启示：一个新的方法专注于解决一个具体的问题可能会带来更大的影响，虽然这个方法具有一般性。

在现在人脸识别的方法中，这两个方法也是follow的人最多的。究其原因，除了其有效性之外，简单是其最大的特点。纵观PAMI历史风云，能经受住时间考验而流传下来的方法，除了有效之外一般都有两个特点其一：1）简单 （PCA, LDA, K-Means, Normalized Cuts etc.）；2）复杂 ，但是解决一个具有一般性而且很难被解决的问题 (在AAM、3d morphable model有深刻影响的Lucas-Kanade算法)。所以如果你的方法一般人都有能力做得到，那就尽量把你的方法做的简单明确。这就是外国人推崇备至的所谓的Ockham's Razor原理 （就个人情感而言，我十分讨厌这个名词）。在这里我要强调一点是，这里说的简单并不是说原理简单，Normalized Cuts就方法本身来说简单，但是原理并不简单；微分几何中的Gauss-Bonnet定理形式非常简单，内涵何其丰富。

在此我想多提两句。由于国内有诸多发论文的方法论，其中一个流传下来的一句话就是：系统做的越简单越好，理论做的越复杂越好。不可否认，这句话有它有道理的地方，但是如果用这句话教育后人，误人子弟矣。

后来出现了许多新的与之类似的方法，就TPAMI上发表的来看，比较有代表性就是 HE Xiaofei 的LPP和 YAN Shuicheng 的MFA。关于这两个方法的评论大家可参看j.liu贴中knato的回帖。
在这里我想谈谈我的个人见解。首先这两个方法的出现有它们的意义。LPP是流形学习中Laplacian Eigenmaps线性化，这样无疑会带动其它流形学习方法在识别问题中的尝试，一个为解决问题找到一个思路，二个为进入寒冬的流形学习找到新的用武之地，虽然这两个都不是上档次的思路，但是潜在影响还是有的。后来 YANG Jian 的UDP就是在LPP号召下在TPAMI上的产物。LPP是非监督方法，所以它的识别性能比LDA好的概率极其微弱。
MFA是基于局部数据关系的监督鉴别方法。它有两个最近临近点数量的参数要调。这两个参数是这个方法双刃剑。参数调的好，MFA会比LDA效果好，调的不好则不行。这样MFA用起来比LDA复杂，这样如果MFA的性能比LDA好的有限，而用起来复杂得多的话，它终将被历史所抛弃。
另外就像j.Liu在他的帖子中说出的一样，这些方法有一定的投机性，比如这两篇文章的试验，他们都把Fisherfaces(PCA+LDA)设为c-1，虽然这是按照原始论文的取法，但是是做过这方面工作的人都知道PCA的主元数目如果取得太大，PCA+LDA的性能会显著降低，在WANG Xiaogang的IJCV上的Random sampling LDA中清楚地给出了图形说明。所以他们论文中给出的实验比较不具可信性。

LPP, UDP, MFA都是我们中国人（至少这些方法发表时还都是）为第一作者发表的方法，个人认为其存在有一定的价值，但它们将是PAMI研究发展中的过眼烟云，无法与PCA,LDA相媲美。



2）LDA奇异性问题


众所周知，LDA是基于求解广义特征值问题（Sb*u=Alpha*Sw*u），所以在实际应用时遇到奇异性的问题，就是Sw矩阵不可逆。在人脸识别中解决这一问题的论文“浩如烟海”。这也说明了LDA的影响力之大。在这一类方法中，也有风格之分。

o. PCA 降维
在Fisherfaces中采用的就是先用PCA降维，再用LDA，这也是现在处理这一问题的一般方法。这里有个比较讽刺的事情。Belhumeur在他的论文里说：PCA actually smears the classes together。那末既然smears the classes together，既然PCA破坏类的结构，那为什莫还要用PCA降维？而且事实证明，即使在Sw可逆的情况下，用PCA features也会增强LDA在人脸识别中的性能。这里只能说明，PCA的作用或是PCA features并不是Belhumeur和其以后follow这样说法的人叙述的那样。PCA虽然简单，但是人们应该对它有个正确的认识，这个以后如果有机会再谈。

a. RDA
至今影响最大最实用的还是基于regularization思想的RDA。其实这个问题不仅仅在人脸识别中才被注意到。很早在统计中就被解决过，RDA发表于1989的Journal of the Americal Statistical Association杂志上，可见其久远。在Sw上加一个扰动项也是解决这一问题的最简单方法。

b.子空间投影
论文最多的也就在这一块。应用knato类似的排列组合方法，令image(Sw)和null(Sw)分别表示Sw的列（像）空间和零空间，则我们可很容易的就列出如下组合方法 （强调：这里却不是提供给大家发论文的方法论，而是以较形象的方式叙述！）
把样本投影到
aa. image(Sb), bb. null(Sw), cc. image(Sw), dd. image(Sw)+null(Sw), ee. image(Sb)+null(Sw) 可并列可串行, ff. image(St)+null(Sw)
以上每一种组合就代表不止一篇论文，在此就不详细列举了。另外，你还可以把random sampling技术加进来，这样就可以不止翻倍。还有，你还可以把同样的技术用到KPCA KLDA (kFA)上，这样又可翻倍。更进一步，你还可以把ICA,LBP, Gabor features等诸如此类的东西和以上子空间混合，...，子子孙孙无穷尽焉。
把这个东西做到极致的是国内的 YANG Jian。另外香港中文大学的 TANG Xiaoou 和他以前的学生 WANG Xiaogang 也做这相关的工作，但是他们做一个idea就是一个，没有灌水之嫌。YANG Jian的工作可以用他在TPAMI上的 KPCA plus LDA 这篇文章来概括，虽然他灌水无数，但就子空间方法而言，他这篇文章还有他发表在国内自动化学报上的那篇长文还是有东西的。如果你想做这一块的工作，值得看一看，是个较为全面的总结。TANG Xiaoou在子空间方面的代表工作（开山之作）就是dual spaces LDA, random sampling (and bagging) LDA, unified subspaces。（在此之后他还有学生一直在做，就不详细列举了。）

我建议想做这一块工作的同学们，要把TANG and YANG的工作烂熟于心，取长补短，相互学习，取其精华，这样可以较为快速而全面地掌握。


c. QR分解
矩阵和数值功底比较好的人，能做得更像模像样。Cheong Hee Park 和 YE Jieping 无疑是这方面的高手。去看看他们在TPAMI,JMLR, 和SIAM的J. Matrix Anal. & Appl上发表的论文可知一二。


d. 相关性
如果Sw可逆，则Sb*u=Alpha*Sw*u可以转化为 inv(Sw)*Sb*u=Alpha*u。那末就可以考察Sw的子空间和Sb子空间的相关性。这方面的代表工作就是Aleix M. Martinez在TPAMI上长文的那个工作。


e. 变商为差
变u'*Sb*u/(u'*Sw*u)为u'*(Sb-Sw)*u。



3）基于图像局部结构的方法

这一类获得广泛认可的方法有Gabor和LBP，另外还有可能有用的SIFT和differential features。
Gabor应用比较早有影响力的代表作就是EBGM。Gabor也是提取用来识别的visual feature的最常用手段。
有无数人因为LBP的极其简单而怀疑它的性能，但是有趣的是最近Ahonen在TPAMI上的短文，就是把LBP应用在人脸识别上，没有任何新的改进，这也说明Reviewer们和editor对这类方法的肯定和鼓励。在非监督feature extraction中，LBP有明显的优势，但是绝对没有达到作者在论文显示的那个水平。在他的论文中，LBP特别weighted LBP效果非常好，这和他们应用的FERET人脸库的人脸crop形式有关。他们应用CSU的椭圆模板来crop人脸，如果应用正方形的模板weighted LBP提高很有限。特别在FRGC Version 2上测试，LBP绝对没有一般监督性的识别方法好。另外这也给我们一个小小启示，就是加个weight其识别性能就能大大提高，这说明什莫问题呢？

另外我不敢苟同j.liu在他文章说的LBP对image blocks大小不敏感是个美丽谎言的说法。首先，有一定的敏感性，这个是要承认的。但是LBP有一个性能稳定的image blocks，并不是人们认为的histogram要符合一定的统计性等等。这个block size的选取比最优的PCA主元数目的选取要容易得多。当然这些都是小问题。

国内有人做Gabor和LBP的结合。当然是值得探索的，但是我个人认为不应该在这两种方法结合上花费太多精力。完全可以用类似形式考虑别的思路。



4) Sparse representation

NMF和NTF都属于sparse representation的方法，都曾被应用在人脸识别中，但效果都非常有限。特别是NTF，属于数学理论上非常优美，但是实际效果很勉强的典型。

另外，Sparse representation (coding) 是一个很有趣也是很有前途的方法，Sparse representation 有很多方式，关键要看你怎莫用、解决怎样的问题。过段时间我们还有机会再谈。



5）Tensor方法

Tensor在人脸识别中至少到现在为止，还非常得不成功。最典型的就是M. Alex O.Vasilescu在ECCV'02上的tensorfaces。他们对于问题的分析和tensor的对应天衣无缝，非常有道理，数学实现上也同样简单，但是自从那个方法发表出来以后基本无人follow。究其原因，个人认为就是把本来简单的问题复杂化，最重要的就是复杂化以后并没有带来该有的益处。

Alex对tensor的应用是flattening high-way tensor。这是一种常见的处理tensor的方法，这样做的好处就是使tensor好处理易于计算。two-way tensorfaces就是我们理解的Eigenfaces。但是同样是tensor，这种tensor和Amnon Shashua的NTF有着本质的区别。NTF是纯正的tensor思想。但是它实现起来过于复杂，又加上原理比Alex的tensor更复杂，所以无人问津。但是不可否认，它们都是数学上十分优美的方法。如果你想学习tensor而又不想枯燥，我推荐你去看这三篇论文（Shashua两篇）。




6）参数模型
参数模型的应用也多种多样，比如HMM, GMM等。这两个都是一般性的建模方法，所以应用也很庞杂，而且在人脸识别中的应用大多是从speech recognition中的方法转化而来，在此就不多谈。有兴趣的同学们可以参看H. Othman在PAMI上的论文和Conrad Sanderson在PR上的论文。

但是在此其中，最简单的是Baback Moghaddam在TPAMI上那个Probabilistic Subspaces的文章，这个文章也是WANG Xiaogang的unified spaces的参考原本。




7) 3D 模型

代表作是Volker Blanz在TPAMI上的那个文章。不过个人十分不看好。



8）Personal Perspectives

a. 基于子空间的方法很难在实际应用中有所用处

b. 基于找图像局部结构的方法更有希望。像EBGM, LBP, SIFT之类可以给我们很多有益的启示。这点和j.liu的观点一致。

c. 把人脸识别中的方法推广开来，应用到一般的分类和统计问题中，这也是人脸识别衍生出来的一大作用。

d. 由于我们国内的特殊研究环境，大家一般都喜欢做简易快的工作，所以人脸识别这一领域出现有华人名字的论文为数可观。其实在某些压力之下这也无可厚非，但是还是希望我们国人在有条件的情况下，不要以发论文为主，多关注于解决问题本身、尽量向推动理论发展的方向努力。我们绝对有这个能力。君不见，NIPS‘06两篇Best student paper被在国外留学的中国人获得，CVPR'07更是又传来喜讯：Best student paper由清华学生获得，这些都是迹象。我们正处于一个意气风发、大有可为的时代。就本人学术水平和资历来说，绝没有资格来说这些话，这只不过是个人的一点心愿和号召而已，同时更是勉励自己。


以上均是dodo个人拙见，囿于本人才疏学浅，难免出现挂一漏万和观点偏颇的情况，还请大家及时批评指正，以免曲彼误人。谢谢

--
Enjoy the world around you. This is life.


※ 修改:・cbir 于 Jul 10 14:00:09 修改本文・[FROM: 211.99.222.*]
※ 来源:・水木社区 newsmth.net・[FROM: 211.99.222.*]

[dahua lin] 我的PhD生活 (转载自水木)

【 以下文字转载自 NLP 讨论区 】
发信人: zibuyu (得之我幸), 信区: NLP
标  题: [dahua lin] 我的PhD生活
发信站: 水木社区 (Sun Mar  8 01:05:06 2009), 站内

一直以来，我在这个blog上写的都是偏重学术的文章。这一期就换一下口味吧，在这里聊聊我日常的生活。

在外面的人看来，MIT这所理工科的殿堂，或多或少有一点神秘的色彩。网上流传着很多关于这个学校的故事，包括Hacker，超负荷的课业，还有各种怪才。这些东西确实真实地存在着，却不是这里校园生活的主流。对于研究生来说，这里的日常生活是相当单调的，这种平凡得乏善可陈的生活和他们特别强调创新的研究工作形成了一种有趣的对比。

PhD的生活方式在很大程度上会受到导师的指导风格的影响。在这里，你问100个不同的PhD，他们会告诉你100种不同的生活方式——但是有一点是共同的，大家的工作都很繁忙。还是说说我自己的吧。在读master的时候，我只需要每过相当长的时间向汤老师汇报一次就行了；而在这里，我必须与三位不同的指导者讨论我的研究。

首先是Eric Grimson，他是我的正式的supervisor。但是，他有另外一重身份——MIT EECS的head，由于管理方面的事务极为繁忙，他在学术界上已经不太活跃了。我一般每个学期会和他有一到两次meeting，对自己的工作做一般性的汇报并且听他的建议，时间不长，通常是30到40分钟。向他做的报告，必须非常简明扼要，再复杂的topic，必须在5分钟内说完，并且要把要点说清楚，这对我来说是一个不小的挑战。然后他会向提出一些问题，并且对以后的大方向提出一些建议。对我来说，我并不期待这个简短的交流过程对我具体的 research有多大的帮助，最主要的是要获得他对我的研究方向的持续支持。

然后是John Fisher，这里的一位Principal Scientist，他是我直接work with的人。和他的接触是相当频繁的，每周会有一次reading group，还有一次一对一的research meeting。在research meeting上，我会很具体的和他讨论我的研究，包括很多细节上的东西。一般来说。他会有很多建议，但是，仅仅是建议，没有要求我必须这样做——事实上有起码50%的建议，会被我当场驳回。不过，这并不影响我和他之间良好的合作，我们都认为这些是正常学术讨论中很自然的事情。

还有就是 Alan Willsky了，LIDS的co-director。他是一个非常渊博的学者，对非常多的学科（信号处理，控制论，统计学习，数学）都着广泛而深刻的了解。每个星期，他和与他有关的学生进行分组讨论，我在其中的一组。我的研究涉及的相当重要的部分——李代数和微分方程，正是他非常熟悉的领域。一方面，我能够从和他的讨论中学到很多东西，事实上，我研究过程中的很多进展都得益于他的启发。另外一方面，他对这个领域太熟悉了，要让这方面的工作得到他的欣赏，是非常困难的。

在这里进行的研究，和我在MSRA或者CUHK做的研究有一个很不一样的地方，我的导师们非常强调一个工作是不是具有开创性的学术价值，而一个算法在实际中work不work，虽然也很重要，但不是放在最核心的地位的。我和他们讨论的绝大部分时间都是理论和方法论上的探讨，至于算法怎么在实验中更好的performance，他们assume是学生自己在实验过程通过各种方式中达到，这些东西如果和理论核心没有特别关系就不会是讨论的主要议题了。这不代表MIT或者CSAIL的全部，不过在我所在的“小环境”里，理论倾向是非常明显的。

在这里，不会有特别的paper或者project的压力，研究是自然地推进的，受会议deadline的影响会有一些（deadline前，如果刚好有一项工作差不多成熟了，需要多花点时间整理成paper），但不是特别明显。研究的最终目标是形成一份有重要影响的PhD thesis，因此，我们不会特别围着CVPR/ICCV/ICML/NIPS之类的会议转。如果留心统计的话，MIT在这些会议上发的文章不会比一个普通的学校多，但是在这里所完成的工作的长远影响远大于一般的学校。

能来到这里的学生，多多少少都希望能在学术上有自己的价值，而不仅仅是毕业后有一份还说得过去的工作——如果仅仅为了这点，用不着来这里，花那么长的时间（在我身边的同学里，5年甚至更长的很普遍的）读一个PhD。不过，现实中总是有着很多的压力和诱惑一点一点地消磨着学术上的理想。一方面，不是在paper-driven的氛围中工作，publication list的增长变得不那么迅速和激动人心；而对于高impact的研究的追求则时常会陷入挫折，推进缓慢。另一方面，faculty的opening逐渐减少和竞争日趋激烈，让前景变得不再是那么明朗。

与之相对比的是来自学术界以外的诱惑，比如工业界，华尔街，和管理咨询公司，它们一直以来都相当青睐从这里出去的学生（无论何种专业），而且有着极富竞争力的待遇。MIT统计了去年的top 5 employer: McKinsey, MIT, Google, Booz Allen, 和Boston Consulting Group，其中三家是顶尖的咨询公司；此外，虽然现在处在金融危机的时代，Morgan Stanley等的著名金融公司的校园招聘还在如常进行，仍旧吸引着很多的学生——而这个学校的学生里面，真正读管理和金融的是比例很小的——这说明了，很多理工科专业的同学去做consultant或者trader了。

这是一种令人困惑或者迷惘的对比。一部分人一直坚持自己的学术理想并取得成功（按照去年的统一，MIT的博士毕业生进入Education的占30%，我相信这里面大部分人并不是去教中学或者小学:-) ），而另外一部分人走进了商业的世界。这没有谁好谁不好的比较，每一次的人生选择都是一种choice——没有标准答案的choice。但是，一旦做出了选择，就意味着你在享受这种选择带给你的一切好的东西的同时，也必须承担所伴随的责任。

直到今天，我依然很执着地认为我会选择学术的道路，这是我内心中觉得最有价值的事情。如果到华尔街去，在为自己创造了财富的同时给世界上的其他人留下了什么——我想今天的局势或多或少表明了这个问题的答案。如果在科学上做出了真正的贡献，那么将给这个世界（至少是所工作的领域）带来进步和改变。一个人一生的价值，不在于他拥有了什么，而在于他创造了什么。

--

※ 修改:・zibuyu 于 Mar  8 01:05:10 2009 修改本文・[FROM: 166.111.135.*]
※ 来源:・水木社区 newsmth.net・[FROM: 166.111.135.*]

what was stumbled upon today

 

Mainly photos:

 

http://prometheus.med.utah.edu/~bwjones/

 

http://www.flickr.com/photos/ethanhein/

 

 

Flickr果然是宝藏

Story of Subrahmanyan Chandrasekhar [转载]

 

转载自： 【杨建邺　发布时间： 2007-03-15 15:09　科学时报】

 

钱德拉塞卡由于第一次突然遭到严重打击而转变研究领域，这一转变居然使他感到受益匪浅，形成了以后不断转变研究领域的特殊风格。虽然不免孤独，却因为每

到一个新的领域它都不可避免的是"新手"，不可能有"傲慢"的可能，只能老老实实从虚心当学生开始。这样倒使得他一生谦逊地对待大自然。这岂不是"塞翁失马，焉知非福"吗？

     我们很少看到印度科学家的传记。这本书使我们有机会了解20世纪30年代前后印度科学家经历的人生历程。所以这本书的翻译出版，可以说填补了一个空白，因此很有价值。仔细看了这本书以后，它给我的震撼和对于科学界曾经发生过的一些不公正事件，有了深入肺腑的了解，而让我感到惊心动魄的是发生在1935年的事件。这件事情几乎可以说决定了钱德拉塞卡将走上"孤独的科学之路"，而且妙不可言的是，它居然塞翁失马得到了一个伟大的启示！要弄清楚其中一些事情和奥妙，还得从1930年讲起。
 
    1930年，钱德拉塞卡带着两篇论文来到了英国剑桥大学。一篇论述的是非相对论性的简并结构，另一篇则论述了相对论简并机制和恒星临界质量的出现。福勒看了这两篇文章，对第一篇他没有什么意见，赞同钱德拉塞卡已取得进展；然而第二篇所说的相对论简并以及由此而生的临界质量，福勒持怀疑态度。福勒把第二篇论文给著名天体物理学家米尔恩看，征求他的意见。米尔恩同福勒一样，也持怀疑态度。
 
    虽然两位教授对钱德拉塞卡的结论持强烈怀疑态度，但钱德拉塞卡通过与他们的讨论和争辩，愈加相信临界质量是狭义相对论和量子统计结合的必然产物。1932年，钱德拉塞卡在《天文物理学杂志》上发表了一篇论文，公开宣布了自己的观点。
 
    1933年，钱德拉塞卡在剑桥大学三一学院获得了哲学博士学位，并被推举为三一学院的研究员。几年来，他与米尔恩已经建立了密切的工作联系和深厚的友谊，他也逐渐熟悉了英国著名的天文学家和物理学家爱丁顿。爱丁顿经常到三一学院来，与钱德拉塞卡一起吃饭，一起讨论问题，爱丁顿几乎了解钱德拉塞卡每天在干什么。
 
    到1934年底，钱德拉塞卡关于白矮星的研究终于顺利完成。他相信他的研究一定具有重大意义，是恒星演化理论中的一个重大突破。他把他的研究成果写成两篇论文，交给了英国皇家天文学会。皇家天文学会作出决定，邀请他在1935年1月的会议上，简单说明自己的研究成果。
 
    会议定于1935年1月11日星期五举行，钱德拉塞卡踌躇满志，自信在星期五下午的发言中，他宣布的重要发现将一鸣惊人。在下午会议上，钱德拉塞卡简短介绍了自己的研究：一颗恒星在烧完了它所有的核燃料之后，将会发生什么情形？如果不考虑相对论性简并，恒星最终都塌缩为白矮星。这正是爱丁顿同意的理论。但是，当人们考虑到相对论简并的时候，任何一颗质量大于1.44M⊙（太阳质量）的恒星在塌缩时，由于巨大的引力超过恒星物质在压缩时产生的简并压力，这颗恒星将经过白矮星阶段继续塌缩，它的直径越变越小，物质密度也越来越大，直到……
 
   "啊，那可是一个很有趣的问题。"他明确地宣称："一颗大质量的恒星不会停留在白矮星阶段，人们应该推测其他的可能性。"
 
    接着，大会主席请爱丁顿讲"相对论性简并"，爱丁顿开始发言了。钱德拉塞卡怀着异常紧张的心情，等待着这位权威的裁定。爱丁顿说：
 
    钱德拉塞卡博士谈到了简并。就此而论，通常使用两种表达："普通的"简并和"相对论性"简并。……我不知道我是否应该逃离这次正在召开的会议，不过我的论文的论点是并不存在像相对论性简并这样的东西！……恒星不得不继续辐射、再辐射和收缩、再收缩，我推测，这样直到它达到几千米的半径为止，此时重力变大，足以抑制住辐射，从而恒星最终能归于平静。……
 
    各种不同的偶然事件也许会介入以拯救恒星，但我希望有比这更多的保护。我认为应有一条自然定律阻止恒星以这种荒谬的方式行动!
 
    钱德拉塞卡惊呆了！怎么爱丁顿从来没有同他讨论过这一点呢？！在那么多的相互讨论中，爱丁顿至少应该表白一下他的观点才对呀！但是，爱丁顿并没有办法驳倒钱德拉塞卡的逻辑和计算，他只是声称，钱德拉塞卡的结果过于"稀奇古怪和荒诞"。这哪里是科学的反驳！
 
    钱德拉塞卡说的这种恒星的最终结局（"直径越变越小，物质密度也越来越大，直到……"），实际上就是现在已被广泛承认的黑洞（black hole），这个名称是30多年后于1969年由美国科学家惠勒正式定下的。但1935年1月11日的那天下午，爱丁顿断然宣布它是绝不可能存在的。他的理由只不过是一种直觉："我认为应有一条自然定律阻止恒星以这种荒谬的方式行动!"
 
    可以想象，1935年1月11日的下午对于钱德拉塞卡来说，真是一个惨淡得可怕的下午。他曾经心疼地回忆过那天下午会议结束后的惨况，他写道：
 
    在会议结束后，每个人走到我面前说"太糟糕了，钱德拉，太糟糕了"。我来参加会议时，本以为我将宣布一个十分重要的发现，结果呢，爱丁顿使我出够了洋相。我心里乱极了。我甚至不知道我是否还要继续我的研究。那天深夜大约1点钟左右我才回到剑桥，我记得我走进了教员休息室，那是人们经常聚会的场所。那时当然空无一人，但炉火仍然在燃烧。我记得我站在炉火前，不断地自言自语地说："世界就是这样结束的，不是砰的一声巨响，而是一声呜咽。"
 
    钱德拉塞卡原本想通过玻尔、泡利等的介入，把这个争论继续下去，但是由于当时物理学家们正忙于建立量子力学而无心介入，钱德拉塞卡的处境变得十分不利，他几乎失去了在英国寻找一个职位的任何机会，人们对爱丁顿的嘲笑记忆极深。没有办法，他只得于1937年来到美国，很幸运的是他在芝加哥大学找到了一个教职。与此同时，钱德拉塞卡决定不与爱丁顿争论，暂时放弃恒星演化的研究，但他坚信他的理论总有出头露面的一天。于是他把他的整个理论推导、计算、公式等统统写进了一本书中，这本书的书名是《恒星结构研究导论》。
 
    写完了这本书以后，他改弦更张，开始研究星体在星系中的几率分布，后来又转而研究天空为什么是蓝颜色的。有趣的是，钱德拉塞卡后来似乎十分满意这种不断转换研究领域的做法，以致他后来又全面地研究了磁场中热流体的行为、旋转物体的稳定性，广义相对论，最后他又从一种全然不同的角度回到了黑洞理论。1983年，他终于因为"对恒星结构和演化过程的研究，特别是因为对白矮星的结构和变化的精确预言"，获得了诺贝尔物理学奖。但这已是他最初提出这种理论后的48年了！
 
    法国著名作家蒙田曾意味隽永地说过：
 
    命运对于我们并无所谓利害，它只供给我们利害的原料和种子，任那比它更强的灵魂随意变转和利用，因为灵魂才是自己的幸与不幸的唯一主宰。
 
    钱德拉塞卡后来的经历，可以说是蒙田说法的一个佐证。1935年1月11日那天下午突然落到钱德拉塞卡头上的严重打击，有可能毁掉一个人的人生；但对于具有"更强的灵魂"的钱德拉塞卡，这一严重的打击却给了他一个千载难逢的机会，使他悟出了一个深刻的道理：为什么科学家到了50岁以后（甚至更早），就基本上不再会有什么创造性了。科学家为什么不能像伟大的文学家、艺术家那样不断地具有创新精神呢？钱德拉塞卡通过自己奇特的经历，找到一个答案，那就是：
 
    "由于没有更恰当的词，我只能说这似乎是有一些科学家对大自然产生某种傲慢的态度。这些科学家有过伟大的洞见，作出过伟大的发现，但他们此后就以为他们的成就，足以说明他们看待科学的特殊方法必然是最正确的。但是科学并不承认这种看法，大自然一次又一次地表明，构成大自然基础的各种真理超越了最强有力的科学家。"
 
钱德拉塞卡举爱丁顿和爱因斯坦为例：
 
    以爱丁顿为例，他是一位科学伟人，但他却认为，必然有一条自然定律阻止一个恒星变为一个黑洞。他为什么会这么说呢？无非是他不喜欢黑洞的想法。但他有什么理由认为自然规律应该是怎样的呢？同样，人们都十分熟悉爱因斯坦的那句不赞成量子力学的话："上帝是不会掷骰子的。"他怎么知道上帝喜欢做什么呢？
 
    而钱德拉塞卡由于第一次突然遭到严重打击而转变研究领域，这一转变居然使他感到受益匪浅，形成了以后不断转变研究领域的特殊风格。虽然不免孤独，却因为每到一个新的领域它都不可避免的是"新手"，不可能有"傲慢"的可能，只能老老实实从虚心当学生开始。这样倒使得他一生谦逊地对待大自然。这岂不是"塞翁失马，焉知非福"吗？这一事例大约会使我们得到很多很多的感受吧？

 

 

2007 news report about Emotiv Corp

Emotiv Ushers New Era of Gaming; Enables Players to Control Games with Their Brains
Posted by: Kevin Hawkins at March 7, 2007 11:50:04 AM
 

Game Developers Conference, San Francisco, CA - March 7, 2007 - Emotiv Systems, the pioneer in brain computer interface technology, today launched the Emotiv Development Kit (EDK) for the electronic games industry. With the EDK, developers will be able to create games that respond to a player's emotions and allow players to control their characters' expressions and manipulate objects using only the power of their brain.

Emotiv's technology represents a scientific breakthrough: it is the only brain computer interface solution that can detect and process both human conscious thoughts and non-conscious emotions, including those represented by brain activity patterns unique to a particular individual. Unlike earlier brain computer interfaces, which only detect a limited number of mental 'states' such as concentration (by identifying when the user is focusing on the screen), Emotiv can process dozens of expressions, gestures and emotions. For the first time, computers will be able to differentiate between thoughts of pushing an object or lifting it; detect a user's smile or win; and respond to emotions such as excitement and calmness.

The EDK is the first product offering from Emotiv, which publicly launched today. It enables game developers to attach dozens of specific thoughts and emotions to many different actions in their game. For example, they can enable players to move an object in a game without the use of a keyboard or joystick, make their character smile when they smile, or require that a player stays calm in order to ensure his or her character remains undiscovered in a stealth game. As a result, developers can create a more interactive, immersive, personal experience than is currently possible.

"The games industry is ripe for a revolution in the way players interact with a game. Current interfaces, such as keyboards and controllers, are relatively basic and non-intuitive and are out-of-keeping with the sophistication levels of today's games and the movement towards more immersive environments," said Nam Do, CEO and co-founder at Emotiv Systems. "Brain computer interfaces dramatically change the way players interact with a game and, as such, have a profound effect on the gaming experience. Developers are looking to this technology to take their games to another level, to differentiate their products and to retain their fans."

Emotiv Development Kit (EDK)
The EDK comprises a headset with multiple sensors for detecting brain activity and a series of application development suites:

_ The Expressiv(tm) suite can identify facial expressions in real-time, allowing developers to create characters that respond to the expressions of the player, such as smiles and winks.

_ The Affectiv(tm) suite measures players' discreet emotional states, allowing a game to respond to the player's emotions, such as excitement or calmness.

_ The Cognitiv(tm) suite detects players' conscious thoughts, enabling them to move or manipulate objects just by thinking about an action, such as push, pull, lift or rotate.

How brain computer interface technology works
The brain is made up of approximately 100 billion nerve cells, which are called neurons. These active neurons cause electrical activity, which can be observed using non-invasive electroencephalography (EEG).

Brain computer interface technology works by observing an individual's electrical brain activity and processing it so that computers can take inputs from the human brain. Human thoughts and emotions can therefore control and influence an application.

About Emotiv Systems
Emotiv Systems is a pioneer in brain computer interface technology. Its focus is on leveraging neuro-technology to create the ultimate interface for the next-generation of man-machine interaction. It does this by evolving the interaction between human beings and electronic devices beyond the limits of conscious interface. Emotiv creates technologies that allow machines to take both conscious and non-conscious inputs directly from your brain. These technologies include a hardware and software platform that can be licensed to commercial software developers and other third parties, as well as a suite of products for consumer applications.

Today, Emotiv is developing solutions specifically for the electronic games industry. In the future, Emotiv's technology has the potential to be applied to numerous industries, including interactive television, accessibility design, market research, medicine, and security.

Founded by four award-winning scientists and technology entrepreneurs, Emotiv is headquartered in San Francisco, CA, and has offices in Sydney, Australia. Investors include Technology Venture Partners, Epicure Capital Partners and the Australian Federal Government. More information is available at www.emotiv.com.

 
 
 previous story  next story 
 "Emotiv Ushers New Era of Gaming; Enables Players to Control Games with Their Brains" Discussion
   
  shengshwi     Member since: 1/16/2007  From: Woodland Hills, CA
 
  Posted - 3/7/2007 3:54:24 PM
This is really interesting. I can definitely see it being a huge market in the future.

 
  User Rating: 1006   |  Rate This User     Report this Post to a Moderator | Link
 

  Tolo     Member since: 3/3/2007  From: Lassa hang, Louangphrabang
 
  Posted - 3/7/2007 10:27:42 PM
cool era! :)

 
  User Rating: 1000   |  Rate This User     Report this Post to a Moderator | Link
 

  Dwiff     Member since: 12/9/2006  From: Eugene, OR
 
  Posted - 3/8/2007 12:24:50 AM
I have been excited about this technology for years, I think it will be a while before a game will outright benefit for the technology though. None-the-less I think it will be one step closer to the next real generation of gaming.

 
  User Rating: 1000   |  Rate This User     Report this Post to a Moderator | Link
 

  Elektordi     Member since: 5/10/2003  From: France
 
  Posted - 3/9/2007 9:54:17 AM
I want one... ;)

 
  User Rating: 1000   |  Rate This User     Report this Post to a Moderator | Link
 

  Rolando     Member since: 6/1/2008  From: Santiago, Santiago de Cuba
 
  Posted - 6/1/2008 1:00:27 PM
It looks really very interesting. Unfortunately the public (or at least the available) demonstrations of Emotiv have the following weak points:
- The available choices on each scene seem to be usually YES or NO (do it or not). A more interesting situation must include several possibilities and show that the user can freely select one of them.
- All the scenes implies the use of hands, arms or face muscles then it is hard to believe that Emotiv is really using brain activity instead of electromyogram.
- The claim of an optimal electrode position is really hard to justify without additional theoretical elements, at least if the goal is brain activity and not electromyogram.
www.electrical-neuroimaging.ch

 
  User Rating: 1000   |  Rate This User     Report this Post to a Moderator | Link
 
[URL: http://www.gamedev.net/community/forums/topic.asp?topic_id=438376&whichpage=1&#2907682]

Article by Professor Allan Snyder FRS

OWNING INNOVATION:  From Idea to Delivery

Home › Publications › Symposia Proceedings › 2002 › Snyder

Academy Symposium, November 2002

Dinner Address Genius, Madness and Innovation Professor Allan Snyder FRS Director, Centre for the Mind/ You know, when you think about it, creativity is an act of rebellion! It is downright subversive! Creativity must, by its very nature, confront conventional wisdom. Distinguished guests, ladies and gentlemen. Creativity, this wondrous expression of our minds, underpins innovation, underpins real genius and even underpins the world's economic growth. And yet, and yet creativity remains largely illusive. We don't really know how it can be nurtured. We don't really know how it manifests itself in some and not in others. Why in some and not in others? I want you to recall "A Beautiful Mind". The recent film about Nobel Prize winner John Nash. John Nash's extraordinary mathematical genius, by his very own admission, came NOT from his conventional training, but rather from his particular form of schizophrenia. Quite simply, John Nash could 'see' patterns and relationships that normal people could not. And, this revelation, resonates with other expressions of genius, from the impressionist artist Gaugin to the Russian novelist Dostoyevsky, and possibly even onto the great Newton. Madness, bi-polar disorder, schizophrenia or whatever, somehow facilitate creativity. Ladies and gentlemen - Why is this so? Why do altered mental states facilitate genius? Answer this question and we will have unlocked one of the mysteries of creativity. Answer this question and we may even allow ordinary people greater access to genius. To be really creative, you must see the world in a new light. Sounds easy, but it is virtually impossible. The disturbing reality is that we don't see what is out there. What we do see is based largely on what we expect to see. What we do see is based largely on what we know. Two people looking at the very same cloud formation can see radically different images. The portrait painter sees a face of dignity, the ultrasound sonographer sees a diseased gall bladder. To me, this says it all. We project out what we know onto everything. Nothing we see is looked at afresh. We are blinded by what we know. We are blinded by our expertise. We are blinded by our mindsets. Have you ever wondered why you can't draw? I mean of course draw without being shown the tricks of how to do so. This really is deeply mysterious, because our brains already possess the information necessary to draw, otherwise we couldn't see at all! For example our brains have algorithms for calculating the shape of an object from subtle shading across its surface. Yet, we are not conscious of this shading. Because, if we were, we would be able to draw without training. When you think about it, when you think about it, this is rather an extraordinary state of affairs. Why does our brain have secret information? But when you think about it, why should we be conscious of subtle shading? Why should we be conscious of any so called backroom deliberations? Isn't it the final decision, the executive decision, and not the intermediate details leading to that decision that is of ultimate importance to us? No wonder subtle details are a secret of the nonconscious mind. But the frightening consequences of this strategy is that we don't see what is actually out there. And, here is the fundamental block to creativity. Is it possible to extricate ourselves from this intrinsic block? Is it possible to access the secrets of the nonconscious mind and see the world the way it really is? Just think of the rich applications if we could actually do that! Now, surprisingly, our approach to accessing the nonconscious is inspired by severely brain damaged people known as autistic savants. You know, like the character Dustin Hoffman played in the Hollywood film, Rain Man. Autistic savants are peculiarly literal. They lack the big picture. They lack executive decision making. Autistic savants would appear to be the exact opposite of the creative mind. But, they display extraordinary skills. Skills which demonstrate that they, unlike us, can access the nonconscious mind. Nadia, a severely mentally retarded girl could draw like Leonardo da Vinci. She did so without any training and from memory. And yet, Nadia was only three years old. She had no language ability and could not even recognise her mother from the nurse. Somehow Nadia could access the mechanisms for vision directly from the raw data of the nonconscious mind. Somehow she could access what is in all of our nonconscious minds. So, a peculiar brain damage affords autistic savants privileged access to the secrets of the nonconscious mind. Privileged access to something that exists in us all but is normally not accessible. Now here's the big idea! Although we do not have access to the nonconscious mind as do savants, is there nonetheless some artificial means to promote this access? Wouldn't it be amazing if, on command, we could switch off the part of the brain that is damaged in savants and allow ordinary people this privileged access to the nonconscious mind? Wouldn't it be amazing if we could just momentarily see the world as it really is? Incredibly, we can! We have actually been able to turn on savant like skills in people by shutting off part of their left brain with magnetic pulses. Incredibly, we can allow anyone access to the unprocessed raw information about the world, information that is normally a secret of the nonconscious mind. And, we do so not by stimulating the brain but rather by turning part of it off! Now this could really have some truly extraordinary applications, especially to learning and problem solving. Because, if we can, through artificial means, allow ordinary people a glimpse of the nonconscious mind, then they too would have an opportunity to see a novel interpretation - a novel way to join up the dots. Then they too would have a greater opportunity for innovation. And, this is precisely what we are attempting right now at our Centre for the Mind. Technological ways to enhance learning and creativity. But how does all this relate to John Nash. How does this relate to mental illness and genius? Well, I wonder if our results with magnetic pulses could help explain why so many geniuses suffer mental disorders? I wonder if, for example, bi-polar disorders, intermittently switch off the left side of the brain, allowing privileged access to raw sensory data of the nonconscious right-brain. This would lead to alternating views of the world. One view that driven by the left brain is consistent with past experience, consistent with what we know. The other view that dictated by the right brain sees the world anew and hence is devoid of familiarity and meaning. Taken together, these alternating views are deeply unsettling because they are unexpected. But, taken together they provide the fertile ingredients for creativity. I am sure you are thinking that there must be other ways than magnetic pulses or mental abnormalities to seeing the world unfiltered through our mindsets. Now, it is very hard to obliterate mindsets, and you need them anyway, so the way to see more of the world is to take on more mindsets. Because the more mindsets you imbue, the more different views you have of the world. So, after mastering one situation, you should go on and master something completely different. And, this strategy emulates the very effects of switching off the left side of the brain because it plunges you into unfamiliar territory. Take Picasso, arguably the most innovative painter of the 20th century. Picasso's four major stylistic shifts were precipitated, in every instance, by an upheaval in his life. He changed everything. He changed his woman. He changed his circle of friends. He changed his house. And, he even changed his dog!!! In every instance, the radical transformations of Picasso's styles, were reflected in the radical transformations of his private life. Take Steven Jobs, the co-founder of Apple computers. After Apple, he started Pixar digital animated films. Animated films? Yes! Returning to Apple years later, he revolutionised the appearance of computers and saved Apple industries. Take my career. My insights about optical fibres for telecommunication were inspired, not from engineering physics, but amazingly from insect eyes. Insect eyes! Working in completely different fields facilitates creativity. Somehow our minds link seemingly disparate concepts into a new synthesis. And, I emphasise, I emphasise the nonconscious nature of this process. Because in the final analysis, true genius is about making nonconscious leaps! Making leaps that explode upon you, seemingly from nowhere. Take the brilliant mathematician Poincare. Poincare's breakthrough solution leaped into his mind, unexpectedly, as he stepped onto a bus, and most importantly, after a lengthy holiday. The problem had incubated in his nonconscious mind. It's this crucial incubation period, the 'let me sleep on it' phenomenon, which links seemingly disparate concepts into a new synthesis. It's this crucial incubation period which facilitates the uprush of innovation and genius. And, it's this crucial incubation period which I believe can be enhanced by new technologies. Ultimately, creativity, this driving force of innovation, is the process of destroying ones own gestalt to build a completely new picture. But as I said, creativity is an act of rebellion! And to initiate a rebellion you must have courage. So, in conclusion remember what the celebrated Sigmund Freud said about his ability to innovate: "I am not really a man of science, I am not an observer, I am not an experimenter, I am not even a thinker. I am nothing but an adventurer - a conquistador - with all the boldness, and the tenacity of that type of being." In other words, in other words, from his own assessment, Freud was not especially skilled or talented. Rather, he had the courage to break the rules and to confront conventional wisdom. Thank you.

Professor Allan Snyder FRS Director Centre for the Mind http://www.centreforthemind.com Centre for the Mind: A joint venture between Australian National University and University of Sydney __________________________________ Centre for the Mind University of Sydney Main Quadrangle (A14) Sydney NSW 2006 Australia Phone: 61 (2) 9351 8531 Fax: 61 (2) 9351 8534 ________________________________ Centre for the Mind (Bdg 59) Australian National University Canberra ACT 0200 Australia Phone: 61 (2) 6125 2626 Fax: 61 (2) 6125 5184   The views expressed in the above article are those of the author(s) and do not necessarily represent the views of the Academy.

Polly Matzinger 从会问问题的女招待到世界闻名的科学家 [转载]

【 以下文字转载自 square 讨论区 】
【 原文由 pico 于 Fri Mar 28 13:02:58 2008 发表 】

◇◇新语丝(www.xys.org)(xys.dxiong.com)(xys.dropin.org)(xys-reader.org)◇◇

Polly Matzinger，从会问问题的女招待到世界闻名的科学家

mdoctor

　　第一次听说Polly的故事，是在研一时免疫学的课堂上。何维老师在讲授
"危险模式"理论时，特意卖了个官子："大家知道提出这个理论的是谁吗？在
国际免疫学界可是一位传奇人物啊！"

　　Polly是一位极具个人魅力的女人。每次学术会议，只要有她的演讲，会场
总是爆满。而且，她曾经为Playboy工作过。当年她做学生时曾向JEM（国际权威
的免疫学杂志）投稿，只署了自己的名字。编辑认为这篇论文只有一个作者并不
可靠，怎么也得把老板的名字也属上吧！文章修回后，增加了一个作者。不久，
论文发表了。后来，Polly的一位同学，也是JEM主编的女儿拿着这篇论文，去找
她老爸说："您知道Polly后边的那个名字是谁吗？"主编说："不是她老板
吗？"女儿哈哈大笑："那是Polly养的小狗（一只阿富汗猎犬）的名字！"此
后，Polly被禁止作为主要作者在JEM上发表文章，直到那位编辑去世。

　　那篇发表在JEM上的论文：Polly Matzinger and Galadriel Mirkwood.
(1978). In a fully H-2 incompatible chimera, T cells of donor origin
can respond to minor histocompatibility antigens in association with
either donor or host H-2 type. Journal of Experimental Medicine, 148,
84-92.

　　就是这么一位传奇的女科学家，在1994年时提出了"危险模式"理论，引发
了免疫学界一场新的"革命"，堪称近10来免疫学最重要的理论突破。随着网络
的不断发展，能从网上搜集到的有关Polly的资料也越来越多。

　　Polly的经历非常特殊，在成为一名科学家之前，做过各种各样的工作，包
括爵士乐手、实验室技术员、训狗员以及Playboy俱乐部的"兔女郎"（Playboy
bunny）等。在她看来，"兔女郎"是一项"伟大的"工作（a great job）。而
在Playboy俱乐部的网站中，Polly也被列入著名的前兔女郎名录：

http://mysite.wanadoo-members.co.uk/explayboybunnies/information/faq2.html
Dr. Polly Matzinger - world renowned immunologist.
Denver Playboy Club, 1969

　　不过，Polly的这些工作都干不了多久，她很快就会感到乏味。1972年她来
到加州Davis做酒吧服务生，同时也使自己有些时间来阅读、写作和从事动物工
作。

　　一天，Davis加州大学的两位教授来到这个酒吧。和往常一样，他们喝着啤
酒，讨论着动物拟态的问题。Polly问："为什么没有动物模仿过臭鼬呢？"
Robert Swampy Schwab教授，也是野生动植物、鱼类、菌类学系主任，竟然被问
得哑口无言。Schwab教授断言，这个"会问问题的女招待"应该成为一个科学家。
于是，他花了9个月的时间，到酒吧给Polly送去各种科学著作，使她相信科学是
永不令人厌倦的工作。1974年，Polly回到了学校，这年她已经27岁。2年后，
Polly获得了迟来学士学位。1979年，获得博士学位，终于开始了她永不觉得厌
倦的科学研究工作。她说，她终生感激那个（引导她走上科学道路的）人（I
owe that man my life）。

　　Polly目前在NIH（国立卫生研究院）的NIAID（国家过敏和传染病研究所）
领导着细胞和分子免疫学实验室的T细胞耐受和记忆研究部门。Polly给自己的实
验室取了一个奇怪的名字："幽灵（GHOST）"。
　　
http://www3.niaid.nih.gov/labs/aboutlabs/lcmi/tCellToleranceMemorySect
ion/matzinger.htm

　　作为一名国际知名的科学家，Polly的文章并不算多，迄今为止在PubMed上
只能检索到不足70篇，但大部分发表在Nature系列（23篇）、Science（4篇）、
JEM（10篇）、Annu Rev（1篇）等顶级学术杂志上。Polly的研究性论文不多，
而大都是综述、评论、讨论，这也是理论免疫学家的一个特点。
　　
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db=pubmed&ter
m=(matzinger%20p[auth])

　　Polly的传奇经历，给我们太多的启示和思考。我们可以赞扬Schwab教授的
伯乐识马，也可以感叹Polly的的特立独行，还可以赞扬那个自由、平等、开放
的学术风气。如果没有Schwab教授，Polly恐怕永远只是芸芸众生中的一员，唯
一值得骄傲的，恐怕就只是曾经做赤戴着可爱的大耳朵和小尾巴的兔女郎。

　　其实，通过网络，我也有幸与Matzinger教授有过几次交流。研究生时，曾
将自己的实验与"危险模式"理论结合，用Email与Matzinger教授讨论过中其中
一些问题，很快就收到了她的详尽的解答和有关实验方向的指导。可惜，由于时
间和条件有限，我没能继续相关研究。

　　前不久，看到Matzinger教授在Nat Immunol上最新的一篇关于"危险模式"
理论的文章，于是再次向她去信，希望获得这篇文章的reprint。第二天，
Matzinger教授便给予了回复。显然，Polly已经不记得我这位3年前曾向她求教
的学生。她说："我对一位外科住院医师能对免疫感兴趣觉得非常惊讶。"同时，
她还附上另一篇几年前发表在Science上的相关的文献，供我参考。虽然实际上
我已经读过阐述"危险模式"理论的大部分文献，包括Science上的那篇，但是，
Matzinger教授此举，着实让我感动。她在用自己的行动，影响着每一个正在或
尚未走上科学之路的渴望科学的年轻人，就像当年Schwab教授带她走上科学之路。

　　Matzinger教授给我的回复：

From: "polly matzinger" <pcm@helix.nih.gov>
To: "**" <**@hotmail.com>
Subject: Re: a reprint (or pdf) of your paper
Date: 2007年8月28日 18:22
hi
i am also sending you another paper about the danger model of immunity
(and the web supplement that went with it), just in case you have not
seen it. i am surprised that a surgery resident is interested in
immunity. good work!i hope you enjoy these, let me know if you can't open
them.
cheers polly

　　科学，对某些人来说，可能只是枯燥乏味的工作，而对另一些人来说，则可
能是永不令人厌倦的快乐的源泉。

　　我想，不管从事什么工作，只要能从中获得快乐，那便是成功。

说明：本文最早于2007-11-03发表于我的博客
http://blog.sina.com.cn/u/1401213573
http://xinguancha.blog.tianya.cn/

(XYS20080327)

◇◇新语丝(www.xys.org)(xys.dxiong.com)(xys.dropin.org)(xys-reader.org)◇◇

Tan Le 报道中所提到的竞争对手 OCZ 公司生产的脑电波信号游戏控制器

脑电波游戏控制器将上市 定价159美元

作者：大文轱辘  2008-03-11

　　【IT168 资讯】再强的游戏鼠标，再好的游戏手柄，恐怕都不能与人脑电波直接控制游戏操作更加吸引人吧，OCZ公司展示过的脑电波系统并不是一项科学发明，而是一种真真切切即将上市的产品！

　　OCZ的"神经脉冲激励者"（NIA）利用一个头带探测脑电波活动，并将其转换为游戏中的各种动作，通过驱动程序定义，可将某种脑电波信号定义为具体的键盘按键或鼠标动作。玩家使用它，不需要动一根指头，就可以在游戏中完成跑、跳、开火的动作。

　　近日OCZ的首席执行官Ryan Petersen在接收采访时说，这套系统的定价为159美元："我们希望每个用户都来使用它，而不是那种高不可攀的概念型产品"。

 

Tan Le on the Cover of Inc. Magazine

 

 Tan Le on the Cover of Inc. Magazine
 December 1, 2008

The scientists at Emotiv have done the impossible: created a brain-wave-reading headset that lets you conjure entire worlds using nothing but your mind -- a breakthrough that could be worth billions. Now comes the hard part.

I'm sitting in a darkened room, attempting to move a large block with nothing but my thoughts. I stare at it intently and imagine myself physically tugging on it, trying to flood my mind with a sense of strain and determination. But the block doesn't budge. I try again, concentrating, concentrating: Move, damn you; I am your master. After a long moment, the block trembles a bit, then slowly skids toward me a few feet before stopping. Encouraged, I mentally bear down until the block resumes its sliding, and this time it keeps going. I'm gripped by the immensity of what I have just accomplished: effecting a change in the world around me without moving a muscle. Well, that's not entirely true. I may have squinted a bit.

This isn't a dream; it's science -- and soon, maybe, a big business. OK, the block was only a virtual one on a computer screen, but that's a nit. The same technology that converted my thoughts into action on the screen someday could be hooked up to a real-life backhoe, robot surgeon, or microwave oven, placing any of those objects at my mental whim. This thought-conversion technology is composed of some extremely sophisticated software and a piece of headgear that looks like a cross between a telephone headset and a skeletal bike helmet. Embedded in the headset are 16 electrodes that press lightly on my scalp, monitoring the electrical signals generated by the 3 pounds of toothpaste-like goo sealed in my skull. The signals are my brain waves, the stuff of thought and emotion. The headset passes the signals to the software, which extracts patterns that can be used to control anything that's run by electronics.

Brain waves usually are monitored in hospitals or research labs, but I'm in a conference room at a company called Emotiv, where a few dozen scientists have developed the gear and software that quite literally read my mind, allowing me to play a sort of video game with nothing but sheer thought. This is not a rough, spare-no-expense research and development prototype of some distant-futuristic product, but rather an upcoming stocking stuffer. For $299, you and yours will very soon be able to vaporize onscreen enemies with an angry thought, have your online characters smile when you smile, and see video games react to your level of excitement.

And that's just for starters. Backed by some impressive partners, Emotiv has a long-range strategy that sounds like a business-school case study from the 22nd century. After enabling us to control video games with our minds, Emotiv intends to let us control most everything else we do on our computers and, after that, what's around our homes. In 10 years or so, according to the company's co-founder Tan Le, we will all go around in a world that will respond to our mental commands. Fed by data wirelessly streaming in from a few freckle-size sensors embedded in your scalp, your stereo will know when you are feeling blue and what sort of music cheers you up. Movies will know when you are getting bored and cut to the action. Car advertisers will know when you are feeling the need for speed. Your doctor will know when you are depressed. Doors will open at your mental command.

Given all this, you might expect that Emotiv would be sitting pretty. But if you think building a mind-reading device is tough, try marketing one. It turns out the old saw about building a better mousetrap doesn't hold in the context of a product most people hesitate to believe is possible and aren't sure they want anything to do with if it is. And that has left Emotiv with a challenge every bit as big as conquering mind reading: figuring out how to present its breakthrough device to the world in a way that will transform it from a slightly scary gadget to the next must-have consumer technology. And Emotiv has to do it while taming persistent hiccups in the system, herding video-game producers into tailoring games to the device, and trying to halt a skidding launch date before competitors -- yes, there are other companies making mind-reading devices -- pick off pieces of the market. "Emotiv faces some crucial decisions it absolutely has to get right," says Stephen Prentice, an analyst at Gartner (NYSE:IT) who has sampled the company's device.

Le admits that such challenges are real. But once consumers give the headset a try, she predicts, a lot of the doubts will themselves be vaporized, and demand will snowball. "We see it becoming a totally ubiquitous device, allowing you to interact in a seamless way with everything else in the world," she says.

That grandiose strategy reflects the intensity and outsize ambitions of Emotiv's founders, and especially of Le. Her entire life has been a string of hard-won, improbable triumphs, and she is loath to lower her standards to anything less than spectacular. Going all in with Emotiv doesn't scare her. "When you start with nothing," she says, "you don't get attached to a lot of things. You end up unafraid to push outside your comfort zone."

It's not entirely true that Le started with nothing. She had a bottle of poison. Her mother kept the bottle and little else on the small, calamitously overcrowded boat on which she, 4-year-old Tan, and Tan's younger sister, grandmother, aunt, and uncle fled the Communist government in South Vietnam in 1981. At first, they had felt lucky to have avoided being captured and jailed. But floating in the South China Sea, they weren't so sure. Pirates were chasing the hapless vessels and picking them off one by one. Hence, the bottle of poison: Tan's mother was determined to grant her children a swift and relatively painless chemical end if their boat should be overtaken. "She didn't tell us about any of the horrible stuff, but she didn't have to," says Le. "You see the fear on people's faces, and you know."

The women and girls were kept in the stifling lower deck all day to make the boat a less appealing target, but they were allowed some fresh air late at night. On the fifth night, when the boat was out of fuel and passengers were down to their final rations, Tan, though she had been warned not to speak while above deck, couldn't resist remarking to her mother on the sudden appearance through the clouds of a wide patch of brilliant stars. "Those aren't stars," her mother gasped. "Those are lights." It was a British tanker steaming alongside. They were rescued and taken to a refugee camp in Malaysia. Three months later, they were given a choice of countries that were accepting refugees. Tan's mother had heard that Australia was a young country with a big future, so that's where the family ended up.

They settled outside of Melbourne, where Le's mother, still in her early 20s, picked vegetables and struggled to learn English in night school while raising her daughters. It's not hard to see why Le grew up with an unshakable belief in her ability to accomplish anything. Her mother, who went on to get a bachelor's and then a master's degree, started a cosmetics business and then a consultancy aimed at facilitating Australian-Vietnamese trade. In 1997, she became mayor of Maribyrnong, a suburb of Melbourne, becoming the first Vietnamese woman to be elected mayor anywhere outside of Vietnam.

Grateful that her family had been able to find a comfortable place in Australian society, Le grew up wanting to help others do the same. At 15, she joined an organization that aids Vietnamese immigrants. Smart, ambitious, and disciplined, she was elected the group's president at 18. Somehow, she also found time to complete her schoolwork and entered Australia's prestigious Monash University at 16. In 1998, Le, then 20, was named Young Australian of the Year, a highly publicized government honor that made her a national celebrity and put her on a speaking circuit, where she hobnobbed with prime ministers, scientists, and international captains of industry. That same year, she graduated with a combined degree in business and law.

Le took a job at one of Melbourne's most prestigious law firms, seeing it as a natural extension of her community service work. But by the time she was 22 and a full-fledged lawyer, she found she couldn't stop thinking about the successful entrepreneurs she had met. In particular, she was captivated by the high-tech moguls, some not much older than herself, who had the ability to forge new types of electronic ties that left people better connected to one another and to the world. "There was a technology revolution going on, and I didn't want to just be a facilitator," Le says. "I wanted to be part of the creating."

In 2000, as her restlessness was peaking, she delivered a speech at the University of Melbourne. Afterward, she was approached by a young Vietnamese student who was studying business and information technology on a scholarship at the nearby Royal Melbourne Institute of Technology. This was Nam Do, like Le newly aspiring to high-tech entrepreneurship. The two hit it off so well that they decided to try to start a company, one that would give Le a chance to make her contribution to the connectivity revolution.

Their idea was for small bar-code scanners that could be built into cell phones so that consumers could aim their phones at products and get a text message back with product information and price comparisons. Telcos weren't interested in the bar-code part but were impressed with the high-speed text-messaging capabilities -- these were pre-American Idol-voting days, and mass text messaging seemed a novel idea. Le and Do sold licenses for the software and stuck in a clause that would allot them a modest-sounding five cents for every message handled by the system. Within a few years, their software was handling 150 million messages a month; you do the math. In 2003, Le and Do sold the company, which they had owned outright. They were 26, rich, and looking for a new -- and bigger -- idea.

Le knew where the pair could grab a little inspiration. A few years earlier on the speaking circuit, she had been at yet another dinner event, feeling a bit overwhelmed as a young Asian woman in a sea of suits, when she spotted another misfit -- a middle-aged man in cargo pants, with wildish hair tucked under a sideways baseball cap. This turned out to be the scientist Allan Snyder, who had a prestigious award of his own to boast about: the Marconi Prize, a near-Nobel-level honor he had been awarded for his role in the development of fiber optics. Snyder and Le got on well and stayed in touch. Le and Do went to dinner at Snyder's home, where he enthralled them with his work on using magnetic fields to stimulate human brains. He went on to bemoan the fact that the computer revolution had shut out emotions, which are, after all, what drive us. The industry had thrived on digital signal processors -- chips and software that could handle images and sounds. What was needed, insisted Snyder, was an emotional signal processor.

The notion rang every bell in Le's head. Snyder was describing a technology breakthrough, an entrepreneurial adventure, and a way to form an entirely new, world-changing type of connection. "We stayed up until 4 in the morning talking about it," Le recalls. "By the time we got together again a few months later, we realized none of us had been able to get the idea out of our heads."

In fact, Snyder had been approached by larger companies about developing his idea. But he liked the idea of starting a company with Le and Do. "There are magical qualities to both of them," he says. "I just had a strong intuition this could work with them in charge." Le brought in another friend: Neil Weste, a prominent Australian chip designer who had sold his last company to Cisco (NASDAQ:CSCO) in 2000 for several billion dollars. Among these four very successful partners, start-up capital would not be a problem for the new company, which they dubbed Emotiv. There was no shortage of strategic vision, either. "We wanted to bring to computers and the Internet all the facial expressions and emotions that are so important in our interactions with each other," Le says.

Emotiv's headquarters looks like that of any Web 2.0 start-up, which is to say it is a cluttered warren with mostly twentysomethings hunched over multiple monitors in San Francisco's South of Market neighborhood. But you have to meet only a few of these laptop lizards to realize that something unusual is going on here. One is an expert on facial expressions. Another has designed high-powered communications software. Yet another has produced best-selling video games. Smoke from a soldering iron wafts from a side room teeming with custom circuit boards. The payroll includes mathematicians as well as an evolutionary biologist.

And then there's the charismatic Le, now 31, who is a bit harder to characterize. She is comfortable shooting the breeze about the fine points of intellectual-property protection, the structure of the human cortex, and the future of the music industry, punctuating all of it frequently with an infectious laugh. But there are also flashes of a less easygoing, sharper-edged Le -- flaring, for example, at the suggestion that Emotiv can be compared with any of the countless start-ups that have set up shop nearby. "They may take on some technology risk in their development, but they know what they want to do is doable," she says. "Here, we're pushing the boundaries of what's possible."

Measuring brain waves, of course, isn't such a big deal. Electroencephalography, or EEG, machines that track the brain's electrical activity at the scalp have been around for the better part of a century. But the best EEG machines cost tens or even hundreds of thousands of dollars -- and for all that, they generally haven't been used for much more than measuring relaxation levels or detecting signs of life.

When they launched Emotiv, the partners figured there was no point in hiring established EEG experts, since the state of the art in EEG machines wasn't even close to what they needed. "We decided that we'd look at the whole landscape of science," says Do, "because there had to be something out there traditional researchers were missing." Ultimately, Emotiv decided to treat emotional signal processing as a sort of math problem that could be solved with clever software. Emotiv opened an office in Sydney and staffed it with mathematicians, digital signal processing experts, and artificial intelligence whizzes. To help keep R&D costs manageable, Emotiv leaned heavily on graduate students willing to work for free in exchange for having some exciting, cutting-edge research on their resumé.

The result was a software program that broke brain waves down into 90,000 components. It was so complex that running a single 10-second brain-wave reading through the program took six computers two days. And sometimes the two-day crunching session would be for naught: The brain-wave readings were so faint that just the electrical activity generated in an eye blink was enough to swamp them. To work well, the software had to learn to filter out the noise. "It was like listening to all the phone conversations in New York at once and trying to pull a few of them out," says Snyder. But the researchers made steady progress, and as they did, Le was quick to file patents; she eventually claimed some 25 that covered a range of processes.

In late 2004, after a day of particularly good progress, the group sensed it was close to being able to read a person's level of excitement in real time. No one went home that evening. Le, Do, and the research team pulled an all nighter; they took turns wearing a standard electrode cap -- sort of like a bathing cap coated inside with gel to improve electrical conductivity -- while watching movies, listening to jokes, arguing, and more, all while a graph on the screen tracked excitement. "By morning, we knew we had it," says Le. "We knew we were going to succeed." Without any champagne on hand and with the bars closed, the team members went to a coffee shop to celebrate, their hair glistening with conductive gel.

By now, I can move that block with ease. I'm ready for a new challenge: making something happen onscreen that has no real-life analog. In this case, I'm to make that same damn block vanish into thin air. What am I supposed to think and feel? Disappear isn't part of my mental repertoire. It's suggested that I stare at the background scene and visualize it without the block. I conjure the image in my mind and focus on making it vivid. The block flickers. I sear the blockless image into my brain, and just like that, the block is gone. Who knew I had the ability to concentrate in such deadly ways? Now for some easier fun. An animated face comes up on the screen, and I'm told to make faces. As I grin, the face grins; it matches my frown, blink, wink, and eyebrow arching. I'm a cartoon! I feel as if the headset is helping me realize fantasies I didn't even know I harbored.

Le and her colleagues were just as tickled when they found they could perform similar feats. But they soon realized they now had a serious decision to make, one they had been putting off while the very feasibility of the project was in play: What do we do with this? Hit the market with an expensive device that would sell in low volume? License the technology to one or more big companies? Or somehow figure out how to bring the costs down enough to sell to a mass audience? The co-founders had been dreamily discussing the possibilities all along, but now they met to formally choose their future. "Nam and I were very excited about the opportunities around licensing, but then Allan said to us, 'We don't want to make money doing this,' " recalls Le. "Nam and I rolled our eyes, thinking that this was typical scientist talk. Then Allan added, 'We want to make a lot of money.' " They all laughed, but the point was clear: They had all seen success in past exploits. Why bother to do this if they weren't going to go for the jackpot? They decided to shoot for the mass market.

The strategy is counterintuitive, to say the least. "The best beachhead strategy for a new technology is one that demonstrates that the technology works, is highly valued by the customer, and gives you a high margin," says Jerome Engel, executive director of the Lester Center for Entrepreneurship and Innovation at the University of California, Berkeley. The transistor, for example, was first brought to market in 1952, when it was used in hearing aids. Customers were grateful rather than finicky, marketing was fairly simple, and the revenue funded expansion into bigger markets. Emotiv, in fact, is working with a wheelchair company to develop a thought-controllable device for those who can't move their body. But that's a sideline. The company's biggest bet remains squarely on consumers -- which Engel finds risky. "If you go for a consumer market first," he says, "you're racing against limited resources, you need to get a lot of partners, and you need to have a very sexy product that delivers exactly what unforgiving customers are looking for. These guys made a choice that carries a huge risk."

The decision to shoot for a mass market immediately led to another: The one market that seemed ripe for a large-scale invasion of innovative interface technology was the video-game industry. "Better and better graphics had reached a point of diminishing returns, while there had been almost no innovation in controllers," Le says. "And gamers tend to be early adopters, making them a good incubator for a new technology." Emotiv opened a new headquarters in San Francisco, placing it close to the heart of the gaming industry, while keeping an R&D team in Sydney.

With a market in mind, Emotiv could now pin down the details of its device. Gamers weren't going to wear a gooey bathing cap, so the team came up with a rigid, relatively unobtrusive, even cool-looking headset able to get an accurate brain-wave reading with 16 gel-free sensors instead of the 128 sticky ones in a standard EEG cap. The headset was augmented with a tiny gyroscope to track head motions and a wireless transmitter to free the wearer of wires. More important, the software's brain-wave-interpreting capabilities were improving by leaps and bounds. The software would eventually be able to differentiate among 30 of what the company characterizes as mental states, roughly divided into three categories -- emotions, facial expressions, and actions. All three types of mental states would be critical: Actions would allow controlling what a character does, facial expressions would convey feelings and intentions to fellow online players, and emotions would allow a game to respond to how a player was feeling. A plot could change when you were bored, a virtual character could appear more often if you found him engaging or threatening, music or lighting could shift to complement your mood.

The technology worked, but it didn't work perfectly for everyone. Some users had more trouble than others sending out consistent, identifiable signals, even after running through a training session. And that, says Le, was a shadow hanging over the future of Emotiv. "If we let something seen as half baked get onto the market, it would be a disaster," she says. "We have an opportunity to revolutionize the way people interact with technology. But we won't get a chance to do that unless we provide the right experience in the beginning."

To make the device easier and more fun to use, Emotiv's team worked furiously with a small video-game-development company called Demiurge Studios in Cambridge, Massachusetts, to embed the technology in a gamelike context. Instead of a boring training session, an Asian sensei walks you through an exotic introduction to your new powers. He gets you to grimace at annoying flying creatures to make them flee, to lift heavy objects, and more. This acclimation process gives the software a chance to record your brain waves and trains you to use them consistently before it throws a series of increasingly difficult challenges at you, such as reconstructing simply via thought a fallen bridge needed for a mystical journey while a fiery sky changes hue in response to your emotional state. Another mini game teaches you to hurl thunderbolts.

The market seemed to break in Emotiv's favor with the success of the Nintendo Wii, which lets users wield game controllers like rackets or steering wheels; the Wii's popularity suggested a real thirst for new sorts of interfaces.

And so, buoyed by early results with test subjects, Emotiv decided to take a chance and unveil a prototype in February 2008, at the closely watched Game Developers Conference in San Francisco. There, new video games and accessories can pick up buzz or sink under the gaming community's disdain.

On the show's opening night, with thousands of attendees and reporters in the audience and video cameras rolling, an Emotiv team member named Zachary Drake attempted to move a cube and more, which by this point was something anyone at Emotiv could do in his or her sleep. But for the first time since the team's big breakthrough, the device, which the team had named the Epoc, simply stopped working. Clearly rattled, Drake gamely tried again and again to work his will on the screen, his face a knot of concentration, his arms reaching out plaintively. For a moment, the crowd was silent. "I think people cringed for us," says Le. Then, the murmuring and snickering began. "Welcome to demo hell, folks," Drake said.

The Emotiv team later learned that a powerful wireless network at the facility had wiped out the connection between the headset and the PC. That the demo might fail had never entered Le's mind, and she just stood there, stunned: "We had done so many dry runs and had never had a problem. I was so shocked. I was speechless." The debacle led to widespread ridicule of the company -- "The Force Is Not Strong With Emotiv's Epoc," "Watch Emotiv's Performance Anxiety," and "The Trade Show and Demo Hall of Shame" were among the headlines on gaming sites.

On the other hand, more than 300 people gave the device a shot at the company's booth, and by almost all accounts, it was a big hit and worked well for virtually everyone who tried it. The company was encouraged enough to set a product launch time frame of the 2008 holiday season. The plan was to sell the headsets through game and electronics retailers, as well as online. Meanwhile, competitors were massing. CyberLearning Technology in San Marcos and OCZ Technology in Sunnyvale, for example, have both developed neural headsets. Hitachi in Japan has poured money into potential mind-reading products, and dozens of universities have made efforts to develop better, cheaper thought processors, any of which could lead to spinoffs. There's even an ambitious project funded by the U.S. military, which hopes to have patrolling soldiers communicating by thought within two decades.

But most notably, there's NeuroSky in San Jose, which has developed a single-electrode game-control headband. NeuroSky's device can detect only two mental states -- attention and meditation. But at a projected $50 or so, it is about one-sixth the price of Emotiv's. And for games, at least, keeping it simple could turn out to be an advantage. "People can use ours right away without training," says Greg Hyver, NeuroSky's vice president of marketing. "You can add on all the features you want to a headset, but if people can't use it right out of the box, they won't use it at all." At a conference in October, Square Enix demonstrated a zombie game that uses the NeuroSky device, and Sega is considering releasing a toy sword or a game based on the technology.

Meanwhile, a company called EmSense is making inroads into the corporate market with a headset that measures the reactions of consumers to games, ads, and other entertainment and marketing creations. EmSense hopes to provide its clients with a detailed, high-tech analysis of what flies with what types of consumers. Coca-Cola used EmSense last year to help fine-tune its Super Bowl advertising decisions. That's a business Emotiv wants to be in on as it moves beyond gaming, and no wonder. Says EmSense CEO Keith Winter: "The market research business is worth billions. It's an ocean. Gaming is a pond."

Le and her teammates have tried these rival systems and remain confident they don't come close to the Epoc's capabilities. Still, Emotiv decided in September to postpone the rollout. Why? Because it can, Le says. She insists Emotiv's technology edge is insurmountable. She also says funding isn't a problem, at least in the near term; the company raised $13.4 million in a round of financing in 2007 -- the Australian government chipped in, along with three venture capital firms. "Trying to make the Christmas time frame just wasn't necessary," she says. "We don't have to risk the whole business trying to meet an early delivery date." Le even suggests that delaying the product could prove to be a smart marketing move. "We want pent-up demand," she says. "We've already got 5,000 preordered through our website. After we deliver a good experience to those early customers, we can talk about making tons of them for next Christmas." The company hasn't announced a new launch date, but Le says the headset will be out in 2009.

Meanwhile, the team continues to think beyond games. The headset already can be used to control most ordinary functions in common software, such as word processing and spreadsheet programs, by taking the place of a mouse -- the cursor simply follows your gaze, and you can think your way into triggering the equivalent of a left or right mouse click. Not only might that be a critical tool for people who may have trouble working a mouse, but it might end up feeling a lot more natural for the rest of us. The technology could be applied to entertainment, Le says, noting that wearing a headset while listening to music or watching a video would allow your computer to track what you like and dislike down to individual choruses or scenes, and start automatically tailoring what it serves you, perhaps via a website -- a sort of brain-powered iTunes that Le hints she would like to see Emotiv own. The headset could help educators who work with children who have autism or attention deficit disorder. Social networking sites could use emotional feedback from the headset to create compatible online gatherings or even assist in matchmaking. Well, maybe. "Love is tricky to identify in brain signals," Le allows. "I'm not sure we know how to tell it apart from lust."

Some powerful partners have come on board. IBM (NYSE:IBM) is working with Emotiv to develop a corporate version of the headset that would allow, for example, virtual conferencing with avatars that represent people's expressions and feelings -- so you would know who was engaged, who was bored, who was laughing at your jokes, and, maybe, who was pretending to laugh. Ketan Paranjape, chief of technical staff at Intel, says the chip giant is interested in enlisting Emotiv's headset to navigate via thought three-dimensional representations of corporate data -- the company featured Emotiv prominently at its annual conference for developers. "We think neural devices will be the next interface," he says.

And that brings us, hypothetically, to the day when we are all wearing Emotiv hair plugs, our thoughts and feelings productively ricocheting through our homes, offices, and, through the Internet, the whole world. That's Le's vision, anyway, and she is almost dismissive of lesser goals. "We don't want to be some niche company providing a specific solution to a specific problem," she says. "We have an opportunity to create an industry that will revolutionize the whole framework of technology."

That may well happen. But first, she has to give the world a chance to move that block mentally, before someone beats her to it.

Contributing editor David H. Freedman is a Boston-based freelance writer.

 [URL: http://www.allianceofceos.com/press/member_news/2008/the_scientists_at_emotiv_have.php]