Tim Vieira

Postdoc@Department of Computer Science, ETHZ - ETH Zurich·美国·OpenReview

研究方向

Structured prediction · Reinforcement learning · Dynamic programming · Natural language parsing · Tokenization · Sampling without replacement · Semiring-weighted logic programming

8.0

Transducing Language Models

ICLR 2026Poster

通讯

8.0

Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo

ICLR 2025Oral

7.8

Language Models over Canonical Byte-Pair Encodings

ICML 2025Poster

一作

7.8

Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Sampling

合作者 (20)

Tim Vieira

Transducing Language Models

Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo

Language Models over Canonical Byte-Pair Encodings

Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Sampling

Better Estimation of the Kullback--Leibler Divergence Between Language Models

The Foundations of Tokenization: Statistical and Computational Concerns

From Language Models over Tokens to Language Models over Characters

Variational Best-of-N Alignment