Все публикации

0:39:57

Tomasz Steifer: Ehrenfeucht-Haussler rank and chain of thought

0:52:06

Kartik Ahuja: On Provable Length and Compositional Generalization

0:50:22

Lekai Chen: LLMs as Probabilistic Minimally Adequate Teachers for DFA Learning

0:32:05

Robert Csordas

0:45:42

Eran Malach: Universal Length Generalization with Turing Programs

0:57:57

Dan Friedman: Representing Rule-based Chatbots with Transformers

0:48:38

Keyon Vafa: Evaluating the World Model Implicit in a Generative Model

0:54:15

Chris Köcher: Hard Attention Transformers on Data Sequences: A Formal Language Theoretic Perspective

0:47:09

Anej Svete: Transformers Can Represent n-gram Language Models

0:48:38

Yash Sarrof: The Expressive Capacity of State Space Models: A Formal Language Perspective

0:38:22

Anton Xue: Logicbreaks: A Framework for Understanding Subversion of Rule-based Inference

0:55:19

Zhiyuan Li: Chain Of Thought Empowers Transformers To Solve Inherently Serial Problems

0:23:24

Yingshan Chang: Language Models Need Inductive Biases to Count Inductively

0:51:24

Alessandro Ronca: On the Expressivity of Recurrent Neural Cascades

1:05:49

Martin Berger: Fast grammar inference on GPUs

0:41:58

Daniel Hsu: Transformers, parallel computation and logarithmic depth

0:22:42

Michaël Rizvi: Simulating Weighted Automata over Sequences and Trees with Transformers

0:36:11

Mark Rofin: Why are Sensitive Functions Hard for Transformers?

0:36:08

Brian DuSell: Stack Attention

0:45:43

Will Merrill: The Illusion of State in State-Space Models

0:38:21

Nur Lan: Bridging the Empirical-Theoretical Gap in Neural Network Formal Language Learning

0:51:07

Dylan Zhang: Transformer-Based Models Are Not Yet Perfect At Learning 2 Emulate Structural Recursion

0:56:32

Giuseppe De Giacomo

0:54:56

Alexander Kozachinskiy: Logical Languages Accepted by Transformer Encoders with Hard Attention