Jinpeng Wang
Publications
VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning
Visual captioning requires models to capture visual content faithfully while minimizing both omission and hallucination. As the dominant paradigm for captioning, MLLMs have achieved strong performance through scaling and high-quality data. Recently, RL has emerged as a key route to driving MLLMs toward higher precision and broader coverage, however, existing reward designs for captioning fail to provide fine-grained and reliable signals for factual verification, limiting their effectiveness. To address this, we propose VCap, a Witness-Adjudicator reward that pairs the reference caption (a witness) with the visual signal (an adjudicator). By explicitly verifying factual consistency between the reference and policy-generated captions grounded in the visual signal, VCap delivers a reward signal with hypergeometric-distribution-level precision for caption quality verification. This design enables effective learning even from imperfect references, facilitating weak-to-strong generalization in RL training. In our experiments, an 8B model trained with VCap outperforms open- and closed-source SOTA models on multiple image and video captioning benchmarks. Human evaluation further confirms its strong alignment with factual correctness. Additionally, VCap improves MLLM perceptual capability, generalizes across tasks, and surpasses best-of-N distillation, challenging prior assumptions about RLVR.
From Verbatim to Gist: Distilling Pyramidal Multimodal Memory via Semantic Information Bottleneck for Long-Horizon Video Agents
While multimodal large language models have demonstrated impressive short-term reasoning, they struggle with long-horizon video understanding due to limited context windows and static memory mechanisms that fail to mirror human cognitive efficiency. Existing paradigms typically fall into two extremes: vision-centric methods that incur high latency and redundancy through dense visual accumulation, or text-centric approaches that suffer from detail loss and hallucination via aggressive captioning. To bridge this gap, we propose MM-Mem, a pyramidal multimodal memory architecture grounded in Fuzzy-Trace Theory. MM-Mem structures memory hierarchically into a Sensory Buffer, Episodic Stream, and Symbolic Schema, enabling the progressive distillation of fine-grained perceptual traces (verbatim) into high-level semantic schemas (gist). Furthermore, to govern the dynamic construction of memory, we derive a Semantic Information Bottleneck objective and introduce SIB-GRPO to optimize the trade-off between memory compression and task-relevant information retention. In inference, we design an entropy-driven top-down memory retrieval strategy, which first tries with the abstract Symbolic Schema and progressively "drills down" to the Sensory Buffer and Episodic Stream under high uncertainty. Extensive experiments across 4 benchmarks confirm the effectiveness of MM-Mem on both offline and streaming tasks, demonstrating robust generalization and validating the effectiveness of cognition-inspired memory organization. Code is available at https://github.com/EliSpectre/MM-Mem.
Towards Efficient Low-rate Image Compression with Frequency-aware Diffusion Prior Refinement
Recent advancements in diffusion-based generative priors have enabled visually plausible image compression at extremely low bit rates. However, existing approaches suffer from slow sampling processes and suboptimal bit allocation due to fragmented training paradigms. In this work, we propose Accelerate \textbf{Diff}usion-based Image Compression via \textbf{C}onsistency Prior \textbf{R}efinement (DiffCR), a novel compression framework for efficient and high-fidelity image reconstruction. At the heart of DiffCR is a Frequency-aware Skip Estimation (FaSE) module that refines the $ε$-prediction prior from a pre-trained latent diffusion model and aligns it with compressed latents at different timesteps via Frequency Decoupling Attention (FDA). Furthermore, a lightweight consistency estimator enables fast \textbf{two-step decoding} by preserving the semantic trajectory of diffusion sampling. Without updating the backbone diffusion model, DiffCR achieves substantial bitrate savings (27.2\% BD-rate (LPIPS) and 65.1\% BD-rate (PSNR)) and over $10\times$ speed-up compared to SOTA diffusion-based compression baselines.