AIConfPaper
← Search

Kunhao Zheng

11 accepted papers

2025

Improving Diversity in Language Models: When Temperature Fails, Change the Loss

Alexandre Verine, Florian Le Bronnec, Kunhao Zheng, Alexandre Allauzen, Yann Chevaleyre, benjamin negrevergne
ICML 2025poster

Increasing diversity in language models is a challenging yet essential objective. A common approach is to raise the decoding temperature. In this work, we investigate this approach through a simplistic yet common case to provide insights into why decreasing temperature can improve quality (Precision…

Cited by 0Source ↗PDF ↗Scholar ↗
2025

Optimizing Language Models for Inference Time Objectives using Reinforcement Learning

Yunhao Tang, Kunhao Zheng, Gabriel Synnaeve, Remi Munos
ICML 2025poster

In this work, we investigate the merits of explicitly optimizing for inference time algorithmic performance during model training. We show how optimizing for inference time performance can improve overall model efficacy. We consider generic inference time objectives with $k$ samples, with focus on p…

Cited by 0Source ↗PDF ↗Scholar ↗
2025

PILAF: Optimal Human Preference Sampling for Reward Modeling

Yunzhen Feng, Ariel Kwiatkowski, Kunhao Zheng, Julia Kempe, Yaqi Duan
ICML 2025poster

As large language models increasingly drive real-world applications, aligning them with human values becomes paramount. Reinforcement Learning from Human Feedback (RLHF) has emerged as a key technique, translating preference data into reward models when oracle human values remain inaccessible. In pr…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

RLEF: Grounding Code LLMs in Execution Feedback with Reinforcement Learning

Jonas Gehring, Kunhao Zheng, Jade Copet, Vegard Mella, Taco Cohen, Gabriel Synnaeve
ICML 2025spotlight

Large language models (LLMs) deployed as agents solve user-specified tasks over multiple steps while keeping the required manual engagement to a minimum. Crucially, such LLMs need to ground their generations in any feedback obtained to reliably achieve the desired outcomes. We propose an end-to-end…

Cited by 25Source ↗PDF ↗Scholar ↗
2025

The KoLMogorov Test: Compression by Code Generation

Ori Yoran, Kunhao Zheng, Fabian Gloeckle, Jonas Gehring, Gabriel Synnaeve, Taco Cohen
ICLR 2025poster

Compression is at the heart of intelligence. A theoretically optimal way to compress any sequence of data is to find the shortest program that outputs that sequence and then halts. However, such Kolmogorov compression is uncomputable, and code generating LLMs struggle to approximate this theoretical…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

What Makes Large Language Models Reason in (Multi-Turn) Code Generation?

Kunhao Zheng, Juliette Decugis, Jonas Gehring, Taco Cohen, benjamin negrevergne, Gabriel Synnaeve
ICLR 2025poster

Prompting techniques such as chain-of-thought have established themselves as a popular vehicle for improving the outputs of large language models (LLMs). For code generation, however, their exact mechanics and efficacy are under-explored using unified metrics and benchmarks. We thus investigate the…

Cited by 7Source ↗PDF ↗Scholar ↗
2023

D4FT: A Deep Learning Approach to Kohn-Sham Density Functional Theory

Tianbo Li, Min Lin, Zheyuan Hu, Kunhao Zheng, Giovanni Vignale, Kenji Kawaguchi +3
ICLR 2023top-25%

Kohn-Sham Density Functional Theory (KS-DFT) has been traditionally solved by the Self-Consistent Field (SCF) method. Behind the SCF loop is the physics intuition of solving a system of non-interactive single-electron wave functions under an effective potential. In this work, we propose a deep learn…

Cited by 10Source ↗PDF ↗Scholar ↗
2023

Distilling Vision-Language Pre-Training To Collaborate With Weakly-Supervised Temporal Action Localization

Chen Ju, Kunhao Zheng, Jinxiang Liu, Peisen Zhao, Ya Zhang, Jianlong Chang +2
CVPR 2023poster

Weakly-supervised temporal action localization (WTAL) learns to detect and classify action instances with only category labels. Most methods widely adopt the off-the-shelf Classification-Based Pre-training (CBP) to generate video features for action localization. However, the different optimization…

Cited by 29Source ↗PDF ↗Scholar ↗
2023

Formal Mathematics Statement Curriculum Learning

Stanislas Polu, Jesse Michael Han, Kunhao Zheng, Mantas Baksys, Igor Babuschkin, Ilya Sutskever
ICLR 2023top-25%

We explore the use of expert iteration in the context of language modeling applied to formal mathematics. We show that at same compute budget, expert iteration, by which we mean proof search interleaved with learning, dramatically outperforms proof search only. We also observe that when applied to a…

Cited by 177Source ↗PDF ↗code ↗Scholar ↗
2022

Prompting Visual-Language Models for Efficient Video Understanding

Chen Ju, Tengda Han, Kunhao Zheng, Ya Zhang, Weidi Xie
ECCV 2022poster

"Image-based visual-language (I-VL) pre-training has shown great success for learning joint visual-textual representations from large-scale web data, revealing remarkable ability for zero-shot generalisation. This paper presents a simple but strong baseline to efficiently adapt the pre-trained I-VL…

Cited by 479Source ↗PDF ↗code ↗Scholar ↗
2022

miniF2F: a cross-system benchmark for formal Olympiad-level mathematics

Kunhao Zheng, Jesse Michael Han, Stanislas Polu
ICLR 2022poster

We present $\textsf{miniF2F}$, a dataset of formal Olympiad-level mathematics problems statements intended to provide a unified cross-system benchmark for neural theorem proving. The $\textsf{miniF2F}$ benchmark currently targets Metamath, Lean, Isabelle (partially) and HOL Light (partially) and con…

Cited by 184Source ↗PDF ↗code ↗Scholar ↗