Main picture

Ph.D. from UCLA
Machine Learning Researcher, Apple AIML
Email: xudehong1996@ucla.edu

Google ScholarLinkedInGitHub

Bio

I am a Machine Learning Reseacher at 🍎Apple AIML, working on LLM post-training and reasoning. I received my Ph.D. from UCLA, advised by Prof. Ying Nian Wu. Previously, I conducted research at Amazon AI.

I am mainly interested in the intersections of language modeling, representation learning, and decision-making. My recent research is focused on building powerful Generative AI models to understand, reason and collaborate with humans. Specifically, my research topics inclide:

  • LLM Post-training and Test-time Reasoning
  • Tool-Augmented LLMs and Agentic Systems
  • Novel Language Model Architectures: Check out our new families of models [LTMs]🧠
  •  

    News

     

    Selected Publications

    * denotes equal contribution.

    Latent Thought Models with Variational Bayes Inference-Time Computation
    , , , , , , , , , ,
    ICML 2025
    We introduce Latent-Thought Language Models (LTMs), a novel language model family that incorporates explicit latent thought vectors. LTMs leverage dual-rate optimization, rapidly updating local latent vectors while gradually refining global decoder parameters. This approach unlocks new scaling dimensions, achieving superior efficiency, perplexity, and zero-shot performance over traditional models. They also exhibit emergent few-shot reasoning, highlighting their potential for advanced language tasks.

    On Conformal Isometry of Grid Cells: Learning Distance-Preserving Position Embedding
    , , , ,
    ICLR 2025 [Oral Presentation (1.8%)]
    This paper explores the conformal isometry hypothesis as a unifying explanation for the hexagonal firing patterns of grid cells. It posits that an animal’s 2D location is encoded as a high-dimensional neural vector lying on a 2D manifold, where local distances in physical space are preserved up to a scaling factor. We are the first paper to provide theortical proof that conformal isometry leads to the emergence of grid cell hexagonality. And we further conduct numerical experiments that such local distance preservation naturally produces the observed hexagonal grid.

    Latent Plan Transformer: Planning as Latent Variable Inference
    , , , , , , , ,
    NeurIPS 2024
    Decision-making via sequence modeling can be viewed as return-conditioned autoregressive behavior cloning. Unaware of their own future behaviors, such models were thought to be susceptible to drifting errors. Decision Transformer alleviates this issue by additionally predicting the return-to-go labels. We propose an unsupervised solution, where a latent variable is first inferred from a target return and then guides the policy throughout the episode, functioning as a plan. Our model discovers improved decisions from suboptimal trajectories.

    Aligning Large Language Models via Fine-grained Supervision
    , , , ,
    ACL 2024
    We propose a method to enhance LLM alignment through fine-grained token-level supervision. Specifically, we ask annotators to minimally edit less preferred responses within the standard reward modeling dataset to make them more favorable, ensuring changes are made only where necessary while retaining most of the original content. The refined dataset is used to train a token-level reward model, which is then used for training our fine-grained token-level Proximal Policy Optimization (PPO) model.

    Diverse and Faithful Knowledge-Grounded Dialogue Generation via Sequential Posterior Inference
    ,
    ICML 2023
    In this paper, we present an end-to-end learning framework, termed Sequential Posterior Inference (SPI), capable of selecting knowledge and generating dialogues by approximately sampling from the posterior distribution. Unlike other methods, SPI does not require the inference network or assume a simple geometry of the posterior distribution. This straightforward and intuitive inference procedure of SPI directly queries the response generation model, allowing for accurate knowledge selection and generation of faithful responses.

     

    Professional Service