The CEU Department of Cognitive Science and the Center for Cognitive Computation (CCC) invites you to the following talk. 

*Please note that this is an online event only, next Tuesday, starting from 3 PM.*

Speaker: Yanli Zhou (New York University)

Title: Compositional Learning of Function Interactions in Humans and Machines

Abstract: Humans are efficient learners of functions - the ability to represent and compose functions emerges early in development. Work in linguistics further suggests that humans are capable of learning much more complex function interactions. Going back to Kiparsky (1968), linguists have cataloged numerous linguistic phenomena covering four

logical patterns for ordering two interactive functions. “Feeding” in a context-free grammar is when one function creates the context for another to operate, and “counterfeeding” is the converse. “Bleeding” occurs when the operation of one function removes the context for another, and “counterbleeding” is its converse. In this project, our aim is to determine the extent to which humans and computational models can learn to compose functions based on the system of interactions. We introduce a learning task that adapts and extends the Piantadosi & Aslin (2016) design, evaluating participants on zero-shot function compositions covering all four interaction types. Our findings indicate that participants can correctly infer the underlying functions based on limited input-output examples; they can also generalize to novel combinations of functions across different conditions. Close examinations of participants’ response patterns reveal a number of potential inductive biases. Furthermore, we demonstrate that a standard sequence-to-sequence transformer model can be trained to complete the same task with high levels of accuracy via meta-learning. Incorporating guidance from human data, our model can learn to reproduce behavioral patterns that mirror the complete and complex way humans perform function compositions.

Time: 3 PM, Tuesday, April 16, 2024

Zoom: Meeting ID: 983 6360 8209 Passcode: 951062

Please, be informed that video/photo recording might take place at the event and the edited version of the video material might be published to communicate or promote CEU PU's activities. Please, find our Privacy Notice here.

Ildikó Varga

Department Coordinator (Budapest) 

Department of Cognitive Science

 

H-1051 Budapest 

Nádor u. 15. FT room 404.

tel: +36-1 327-3000 2941

http://www.ceu.edu

http://cognitivescience.ceu.edu

From: Ildiko Zsoka Varga
Sent: Tuesday, April 9, 2024 12:00 PM
To: 'talks@cogsci.ceu.edu (talks@cogsci.ceu.edu)' <talks@cogsci.ceu.edu>
Subject: CCC Colloquium: Yanli Zhou, Tuesday,16 April, 3 PM

 

Dear All, 

 

The CEU Department of Cognitive Science and the Center for Cognitive Computation (CCC) invites you to the following talk. 

*Please note that this is an online event only, next Tuesday, starting from 3 PM.*

Speaker: Yanli Zhou (New York University)

Title: Compositional Learning of Function Interactions in Humans and Machines

Abstract: Humans are efficient learners of functions - the ability to represent and compose functions emerges early in development. Work in linguistics further suggests that humans are capable of learning much more complex function interactions. Going back to Kiparsky (1968), linguists have cataloged numerous linguistic phenomena covering four

logical patterns for ordering two interactive functions. “Feeding” in a context-free grammar is when one function creates the context for another to operate, and “counterfeeding” is the converse. “Bleeding” occurs when the operation of one function removes the context for another, and “counterbleeding” is its converse. In this project, our aim is to determine the extent to which humans and computational models can learn to compose functions based on the system of interactions. We introduce a learning task that adapts and extends the Piantadosi & Aslin (2016) design, evaluating participants on zero-shot function compositions covering all four interaction types. Our findings indicate that participants can correctly infer the underlying functions based on limited input-output examples; they can also generalize to novel combinations of functions across different conditions. Close examinations of participants’ response patterns reveal a number of potential inductive biases. Furthermore, we demonstrate that a standard sequence-to-sequence transformer model can be trained to complete the same task with high levels of accuracy via meta-learning. Incorporating guidance from human data, our model can learn to reproduce behavioral patterns that mirror the complete and complex way humans perform function compositions.

Time: 3 PM, Tuesday, April 16, 2024

Zoom: Meeting ID: 983 6360 8209 Passcode: 951062

Please, be informed that video/photo recording might take place at the event and the edited version of the video material might be published to communicate or promote CEU PU's activities. Please, find our Privacy Notice here.

Best regards,