Profile

Cover photo
Gary Marcus
2,562 followers|49,186 views
AboutPostsPhotosVideos

Stream

Gary Marcus

Shared publicly  - 
 
Has Big Data run amok? An oped I wrote for the NYT, with Ernie Davis.
It’s a valuable tool for analysis, but don’t believe all the hype.
6
6
Vadim Nasardinov's profile photoFacundo Jorge's profile photo
Add a comment...

Gary Marcus

Shared publicly  - 
 
 
LCSL, IIT-@MIT_CSAIL Brains, Minds & Machines Seminar Series '13-'14 | Apr. 22 in Boston, @garymarcus on computational diversity...
1
Add a comment...

Gary Marcus

Shared publicly  - 
 
Is the brain a kind of computer? Why or why not? Discuss. I propose that the brain is not a serial von Neumann machine,, but that it is a computer in the sense that it accepts inputs, and transforms those inputs into outputs, in systematic ways. 
3
1
John Kubie's profile photoZak Wardell's profile photoGary Marcus's profile photoJona Sassenhagen's profile photo
6 comments
 
You'll notice that if you try to spell out "decision", "define", "self-aware", your argument doesn't demarcate computers and brains anymore. See self-modifying code, cognitive quasi-modularism, ... Many philosophers of mind will actually hit you if you say brains are self aware.

The question thrives entirely based on two different approaches to conceptualising it: if you go by the prototype, your answer is "no", if you go by the definition, it's "yes". Of course, Gary Marcus is well familiar with prototype theory (Rosch) and neuronal computations (his own work), so maybe this is a psychological experiment :)
Add a comment...

Gary Marcus

Shared publicly  - 
 
brain implants & how they will eventually change society; my latest, with Christof Koch
4
3
William Payne's profile photoEugenio Culurciello's profile photo
Add a comment...

Gary Marcus

Shared publicly  - 
 
Artificial intelligence on CBS!  and thoughtful comments by +Yann LeCun 
 
A short CBS segment with +Gary Marcus commenting on the future of AI. Well, it's more about the potential dangers of AI for society than about the progress of AI.

In addition to the potential dangers, it would have been nice to also mention the numerous potential benefits of AI: transportation safety, medicine and health care, manufacturing, etc. Some economists have said that AI will be the engine of the next industrial/economic revolution. 

The potential dangers of AI should certainly be taken seriously, even if many of them are far off in the future. And we should have AI ethics committees to discuss and control the possible effects of AI on society, the same way we currently have committees on bioethics, nuclear technology, climate change, public health, etc. 

Every technology can be used for good things and bad things. Common decency and working towards the common good is what drives us to use it for good things. Laws, treaties and regulations is what prevents society from using it for (too many) bad things. 

As Gary said (and unlike James Barrat's remark to the contrary), we are still quite far from having "truly" intelligent machines (whatever that means), and from machines with "common sense". But we are making quick progress in visual and auditory perception and with language understanding.
3
2
Daniel EmoSPARK's profile photoDaniel Panichelli's profile photo
Add a comment...
Have him in circles
2,562 people

Gary Marcus

Discussion  - 
 
 
LCSL, IIT-@MIT_CSAIL Brains, Minds & Machines Seminar Series '13-'14 | Apr. 22 in Boston, @garymarcus on computational diversity...
1
Add a comment...

Gary Marcus

Discussion  - 
 
 
Is the brain a kind of computer? Why or why not? Discuss. I propose that the brain is not a serial von Neumann machine,, but that it is a computer in the sense that it accepts inputs, and transforms those inputs into outputs, in systematic ways. 
1
Mark Waser's profile photoLotzi Bölöni's profile photoJoshua Augustus Bacigalupi's profile photoDeepak J. Nath's profile photo
18 comments
 
+Joshua Augustus Bacigalupi  I understand what you are saying. But you should look at deep neural networks. They do computation quite differently that normal computers and infact does the exact things you are mentioning for image recognition and voice recognition.
Add a comment...

Gary Marcus

Discussion  - 
3
Samuel Kenyon's profile photoNikolai Varankine's profile photoBen Bogart's profile photoJoshua Augustus Bacigalupi's profile photo
6 comments
 
By chance I opened up a random science fiction anthology and read a story from 1951 by Bernard Wolfe called "Self Portrait" about cyberneticists making advanced prosthetics (and another team working on a chess-based military strategy computer). It was amazingly correct in regards to people thinking that robotic prosthetics would be achievable quickly, and then they find out that the kinesthetics/mechanical aspect is extremely far easier than the nervous system interaction/feedback aspect, and in the story the amputee test subject ends up just controlling the prosthetic leg with some push buttons. It's almost exactly what happened with the advanced prosthetic arms made during the past 10 years (like the "Luke" DEKA arm).
Add a comment...

Gary Marcus

Shared publicly  - 
 
 
Today's edition of All Things Considered on NPR has a very nice segment on deep learning in which I am interviewed, and my dear colleague +Gary Marcus  comments on AI, deep learning and such.

Featured in the segment is the work of my just-graduated PhD student +Pierre Sermanet on recognizing cats from dogs (his winning entry in a recent Kaggle competition).

http://www.npr.org/blogs/alltechconsidered/2014/02/20/280232074/deep-learning-teaching-computers-to-tell-things-apart
2
1
Deepak J. Nath's profile photo
Add a comment...
16
12
Steve Esterly's profile photoYasin Gülener's profile photoTim Osmar's profile photoChristoph Nahr's profile photo
4 comments
 
+Gary Marcus, i enjoyed to read your article after reading the nyt's piece from last weekend stating that circuits working like the brain are already being produced. i think that one was full of wishful thinking and i tend to agree with your analysis. however, i agree with +Tau-Mu Yi and +Jochen Fromm that we do understand a bit about the brain. i am almost finishing this book http://www.amazon.com/Thinking-Science-Decision-Making-Problem-Solving-Prediction-ebook/dp/B00BATNLHO/ref=sr_1_1?ie=UTF8&qid=1388718458&sr=8-1 which has been very informative and my team and i have been working with some predictive models that, thanks to the computational power now available at a much cheaper price point than before, will begin to deliver brain-like answers for some problems. we don't need to simulate the whole brain to get brain-like results. keep up with the good writing, thank you!
Add a comment...
People
Have him in circles
2,562 people
Basic Information
Gender
Male
Story
Tagline
author of Guitar Zero,, frequent blogger for The New Yorker
Introduction
Author of Guitar Zero and Kluge, Professor at NYU. Blogs at The New Yorker.