You are here

Reverse engineering human common sense

Date and Time: 
Thursday, January 21, 2016. 04:00 PM - 05:30 PM
Meeting Location: 
Building 260, Room 113
Meeting Description: 
Speaker:
Joshua Tenenbaum is Professor of Brain and Cognitive Sciences at MIT.
 
Meeting description:
What is the essence of human intelligence — what makes any human child smarter than any artificial intelligence system that has ever been built?  Our work aims to answer this question in computational terms, with a precision sufficient to engineer this cognitive capacity in a machine and to generate behavioral predictions for both behavioral and neural experiments.  We are inspired by recent developmental psychology studying the basic commonsense knowledge of young infants and the basic learning mechanisms by which children grow this knowledge over their first years of life.  What we call the ‘’common sense core’’ is a set of integrated cognitive capacities for scene perception, physical and social reasoning, and rapid (often “one-shot”) learning of rich abstract knowledge from sparse data.  These capacities are present prior to language and provide the building blocks of language understanding and production, and action planning.  I will show how we are begin
ning to capture these capacities computationally using the techniques of *probabilistic programs* and *program induction*, embedded in a broadly Bayesian framework for inference under uncertainty.  These methods could also lay the foundations for more human-like approaches to artificial intelligence and natural language understanding, and computational neuroscience approaches that connect more deeply with the core of human cognition.

Workshops Calendar

S M T W T F S
 
 
 
 
 
1
 
2
 
3
 
4
 
5
 
6
 
7
 
8
 
9
 
10
 
11
 
12
 
13
 
14
 
15
 
16
 
17
 
18
 
19
 
20
 
21
 
22
 
23
 
24
 
25
 
26
 
27
 
28
 
29
 
30