Related papers: Internal Flow Signatures for Self-Checking and Ref…

Sparse Identification of Nonlinear Dynamics (SINDy) is a powerful method for discovering parsimonious governing equations from data, but it often requires expert tuning of candidate libraries. We propose an LLM-aided SINDy pipeline that…

Reasoning is a fundamental substrate for solving novel and complex problems. Deliberate efforts in learning and developing frameworks around System 2 reasoning have made great strides, yet problems of sufficient complexity remain largely…

This project develops a self correcting framework for large language models (LLMs) that detects and mitigates hallucinations during multi-step reasoning. Rather than relying solely on final answer correctness, our approach leverages fine…

Breaking down a problem into intermediate steps has demonstrated impressive performance in Large Language Model (LLM) reasoning. However, the growth of the reasoning chain introduces uncertainty and error accumulation, making it challenging…

Security and distributed infrastructure are two of the most common requirements for big data software. But the security features of the big data platforms are still premature. It is critical to identify, modify, test and execute some of the…

Network Intrusion Detection Systems (NIDS) have progressively shifted from signature-based techniques toward machine learning and, more recently, deep learning methods. Meanwhile, the widespread adoption of encryption has reduced payload…

Unsupervised Reinforcement Learning from Internal Feedback (RLIF) has emerged as a promising paradigm for eliciting the latent capabilities of Large Language Models (LLMs) without external supervision. However, current methods rely on…

Simulink/Stateflow charts are widely used in industry for the specification of control systems, which are often safety-critical. This suggests a need for a formal treatment of such models. In previous work, we have proposed a technique for…

Binary vulnerability analysis is increasingly performed by LLM-based agents in an iterative, multi-pass manner, with the model as the core decision-maker. However, how such systems organize exploration over hundreds of reasoning steps…

Code review is a critical practice in software engineering, yet the growing scale and frequency of code patches in modern projects, together with the widespread adoption of AI code assistants, make manual review increasingly challenging.…

Ensuring that large language models (LLMs) remain both helpful and harmless poses a significant challenge: fine-tuning on repetitive safety datasets, where unsafe prompts are paired with standard refusal templates, often leads to false…

Recent IoT applications gradually adapt more complicated end systems with commodity software. Ensuring the runtime integrity of these software is a challenging task for the remote controller or cloud services. Popular enforcement is the…

We propose ReinFlow, a simple yet effective online reinforcement learning (RL) framework that fine-tunes a family of flow matching policies for continuous robotic control. Derived from rigorous RL theory, ReinFlow injects learnable noise…

Vision-Language Models (VLMs) have recently shown promising advancements in sequential decision-making tasks through task-specific fine-tuning. However, common fine-tuning methods, such as Supervised Fine-Tuning (SFT) and Reinforcement…

Diffusion-based unsupervised image registration has been explored for cardiac cine MR, but expensive multi-step inference limits practical use. We propose FlowReg, a flow-matching framework in displacement field space that achieves strong…

We propose V2Flow, a novel tokenizer that produces discrete visual tokens capable of high-fidelity reconstruction, while ensuring structural and latent distribution alignment with the vocabulary space of large language models (LLMs).…

Intrinsic self-correct was a method that instructed large language models (LLMs) to verify and correct their responses without external feedback. Unfortunately, the study concluded that the LLMs could not self-correct reasoning yet. We find…

How do reasoning models verify their own answers? We study this question by training a model using DeepSeek R1's recipe on the CountDown task. We leverage the fact that preference tuning leads to mode collapse, yielding a model that always…

LLM watermarks allow tracing AI-generated texts by inserting a detectable signal into their generated content. Recent works have proposed a wide range of watermarking algorithms, each with distinct designs, usually built using a bottom-up…

Learning permutations is fundamental to sorting, ranking, and matching, but existing differentiable methods based on entropy-regularized Sinkhorn produce a single softened solution and collapse under ambiguity. We present PermFlow, a…