2017年8月2日星期三

Big data or big 理论? Gallant and Hickok discuss.

几个月前,杰克·加兰特(Jack Gallant)发表了一些引起我注意的评论。这是一个:

问题是大多数MRI经验。设计基于行为。心理,这是一个糟糕的框架。

I disagreed on grounds that psychology provides 在 least one part of a 理论 that constrains MRI research.

这是杰克的另一条推文:

In 理论, 理论 is great. In practice currently, only data-driven models accurately predict brain activity under naturalistic conditions.

为此,我回应:


杰克,你的残局是什么?想要预测大脑活动吗?还是您想了解大脑如何计算思维?
杰克的答案:
Science's end game is always an elegant, predictive 理论. But complex systems often require a data-driven middle game.
然后我和杰克之间的讨论脱机了。 经过几次交流,在我看来我们似乎有点融合,并且肯定在阐明我们的立场。 我觉得讨论对其他人来说很有趣,并问杰克是否可以在这里发布。 他同意了,所以你去这里! Comments welcome!

JG:

我是这样看的:(1)可以很好地预测但您不知道为什么的计算模型,或者(2)可以理解但没有预测的模型?我想说显然(1)是首选。如果您拥有一个可以预测但无法理解的计算机计算模型,则可以研究该模型而不是大脑。当然,这会容易得多,因为您可以对模型进行的实验数量不受限制。相反,如果您有(2),则可能会陷入无关紧要的局部最小值中,从而完全浪费时间。

Understanding (i.e., a low-dimensional explanation that accurately predicts) is obviously the ultimate goal of science, but you may not get there in the most straightforward path (i.e. through 理论-driven approaches).

GH:

Hi Jack, yes a nice predictive model is great. My point though is about what we are trying to predict. Your statement makes it sound like all we are trying to predict is physiology. That's fine for a physiologist, but the point about studying the brain is that it is a system for controlling behavior. We therefore need good data and good theories of brain, behavior and their relation. I think you agree but many of your tweets give an anti 理论 anti cog sci impression.

JG:

嘿,格雷格(Hreg Greg)信不信由你,我认为我们同意除优先事项之外的所有其他事项。因此,让我总结一下我认为问题出在哪里,您就可以纠正我。我们俩都认为大脑是控制行为的某种多肉计算机。我们俩都认为行为最终是最有趣的事情。 However, you seem to think that 理论 is really useful and important AT THIS TIME for studying the brain and the brain-behavior relationship. And I do not. My reasons for this are that (1) our understanding about how a system like the brain might compute are really poor, because we don't really understand distributed nonlinear dynamical systems like the brain, (2) we are severely data-limited because our brain measurement technology is pretty poor, and (3) other 在 tempts to use 理论 to predict computational principles of brain function beyond the most peripheral stages of sensory processing have largely failed. 仅以视觉为例。除了V1和MT,没有很好的视觉功能理论。在这些领域之外运行良好的所有模型都是数据驱动的,而不是理论上的。 And note that the SAME PROBLEM arises in computer vision, and in NLP. The models that actually WORK in computer vision are neural nets, which are basically a data-driven universal mapping function. And the models that actually WORK in NLP are neural nets. That is why all of the engineering people have (temporarily) abandoned 理论 in favor of nets. Now ultimately of course we're going to have to take these data driven models and extract their principles of computation. But that is a very different problem from starting with overly-strong 理论 and then ending up in a local minimum.

 GH:

 谢谢你您的观点现在更加清晰。我们所谈论的事情略有不同,或者至少我们正在采取不同的方法。您似乎正在采取一种工程方法,其下一个目标是试图弄清楚预测神经活动的方法,并且希望一旦完成,我们就可以得出一些“计算原理”。我担心的是,如果不做一些认真的理论工作就可以得出数据驱动的结果,那么就没有简单的方法可以从工程方法学到原理。 因此,尽管我喜欢这种工程方法并认为这是值得追求的,但我认为如果不并行进行理论工作,它就无法回答我们的问题。

8条评论:

威廉·马汀说过...

I study people who are starting to acquire their first language late, and who are clearly quite different from those that have acquired it from birth. Theory tells me that these people can acquire words but not a grammar. What can the modeling approach possibly do to help me understand what is different about these people? 我不't see any use in modeling unless it includes a 理论 (i.e., claims about ontology).

未知说过...

有趣的讨论。我不得不说,我倾向于就这一问题与JG达成共识。传统上,我的领域(无视障碍症)遭受的数据太少,理论主张也非常宏大。确实,这使我们(几乎)无处了解失语症的相关机制(认知或生理),更重要的是,它帮助人们从失语症中恢复过来。目前,我确实认为我们需要更大的数据集,这最终将帮助我们提出更强大的理论,以对行为进行预测并为康复提供信息。

-朱利叶斯

格雷格希科克说过...

我不'认为没有人在争论大数据的有用性,我同意,过去有关失语症的理论研究不足。插口's position, though, is that there is little role for 理论 在 this stage and we should focus exclusively on building data-driven models that predict new sets of data, even if we don't understand why it does the predicting. My position is we need both big data and big 理论 working together in lockstep. That's为什么将我带入您的C-Star项目,对了朱利叶斯?看看我们正在取得多少进展! ;)

未知说过...

在实践中,没有模型可以完全预测。某些错误表明您的模型缺少该现象的基本属性'应该建模,但其他类型的错误却没有't. There's no way to identify those kinds without a 理论 of the task (what is a "fundamental property"). In other words, you need 理论 to construct the test set you use to evaluate your model.

未知说过...

OK, I guess I better temper my earlier enthusiasm since I strongly believe some theoretical models are very useful/important, especially for understanding normal function. However, there are clearly areas where we severely lack data. E.g. look 在 agrammatism in aphasia. There are multiple theories that have 在 tempted to explain this impairment, many without much (any?) data support. This is a good example where we probably need lots more data to formulate an informed 理论. -朱利叶斯

尼尔斯·詹森(Niels Janssen)说过...

Thanks this is an interesting debate. However, I am not sure that this is entirely satisfactory. It seems to me that both JG and GH agree that there should be a 理论 of behavior, but (1) how to get to that 理论, and (2) what the character of that 理论 should be is left unclear in my opinion. So, with respect to the first point, I think I read from this that JG endorses a more bottom up approach (generate lots of brain data, let those data drive 理论 development), while GH endorses a more top down approach (generate hypotheses, try to confirm those with targeted data). Is that a correct interpretation?

With respect to the second point, it seems that JG likes to see a more biological 理论 that explains behavior, while GH likes to see a psychology/cognitive 理论 in between brain and behavior. Is that a correct interpretation?

最后一点是,所有这些自下而上的自上而下的讨论都强烈地使我想起了过去的连接主义辩论。

格雷格希科克说过...

Julius -- Yes, agrammatism is the poster child for more 理论 than data. Let's fix that!

尼尔斯(Niels)-我认为您的解释大致正确,您提出了关于"theory"意思是这里。我相信我们需要三种理论。一是关于大脑如何工作(即,它如何计算东西),一是关于大脑的工作(即,它需要执行的任务以及如何表示和转换信息以完成这些任务),一是关于大脑之间前两个(即,大脑实际上是如何计算思想的's processes). For the latter, I think that an important key will be in the neural architectures. True, the debate is very similar the old connectionism debate. There is no right way, in my view; we need all the ways to make real progress, which is why I bristle hearing statements to the effect that we need to focus on bottom up big data and hope a 理论 emerges from it.

拉尔夫说过...

嗨,我前段时间试图对此发表评论,但后来。提交期间发生了评论,评论丢失了。但是,由于我一遍又一遍地回到这个问题上,因此我决定再次尝试...
可能有许多读者认为,我在这两个职位之间陷入困境-专业人士与(让'这样称呼)反理论。在过去的几年中'我已经阅读了很多有关新的数据驱动方法,预测,机器学习,特别是深度学习的文章。显然,我感兴趣的许多领域都取得了很多成功,尤其是在理解自然图像和语音方面。这些方法在理论上通常是很不了解的(例如:Ng'关于语音学是由语言学家组成的陈述-不确定他是否真的说过,但它能捕捉到精神),但同时也具有理论性(例如受神经网络的启发)-至少我觉得他们对理论有很大的潜力(例如克里格斯科特和加兰特'的工作最终可能会提高心理素质。理论/计算问题....)。但是,令我困扰的是,我觉得自己没有足够的能力在理论层面上讨论这些新颖的发展,因此无法将它们与我的知识相结合,更不用说利用它们来推进理论发展了。实际上,这些天花在方法上的时间太多了-至少我-几乎忘了理论是什么(以及嵌齿轮神经中有哪些...)。一世'已找到,例如这本书"系统神经科学中的23个理论问题"非常有帮助,而且您的关于镜像神经元(神话)的书。当然,也有一些老经典,例如Sejnowski和Churchland,Marr等。但是,标准教科书大多'不能过多地强调理论(例如Gazzaniga,Kandel'非常生物起源)。受社会科学启发的有关理论建构的书籍基本上没有'不能给我我希望得到的东西。'我的问题是:(社会,认知,情感)神经科学领域的哪些书籍是学习如何理论化或正确教授理论的构建和分析的好资源?任何阅读技巧将不胜感激..拉尔夫