In visual question answering (VQA) context, users often pose ambiguous questions to visual language models (VLMs) due to varying expression habits.Existing research addresses such ambiguities primarily by rephrasing questions.These approaches neglect the inherently interactive nature of user interactions with VLMs, where ambiguities can be clarified through user feedback.However, research on interactive clarification faces two major challenges: (1) Benchmarks are absent to assess VLMs' capacity for resolving ambiguities through interaction; (2) VLMs are trained to prefer answering rather than asking, preventing them from seeking clarification.To overcome these challenges, we introduce ClearVQA benchmark 1 , which targets three common categories of ambiguity in VQA context, and encompasses various VQA scenarios.Furthermore, we propose an automated pipeline to generate ambiguity-clarification question pairs.Experimental results demonstrate that training based on the automated generated data enables VLMs to ask reasonable clarification questions, thereby generating more accurate and specific answers based on user feedback.

Current VLM-based VQA methods often process entire images, leading to excessive visual tokens that include redundant information irrelevant to the posed question.This abundance of unnecessary image details creates numerous visual tokens, drastically increasing memory and computational requirements in VLMs.To address this, we propose Contextual Region-Oriented Visual Token Pruning (CROP 1 ), a novel framework to compress visual tokens through a two-step process: Localization and Pruning.Specifically, CROP first employs an efficient model to identify the contextual region relevant to the input query.Subsequently, two distinct strategies are introduced for pruning: (1) Pre-LLM Compression (PLC), which adaptively compresses different image regions with varying ratios, and (2) Inner-LLM Pruning (ILP), a training-free method that prunes tokens within early LLM layers guided by the identified contextual region.Extensive experiments on a wide range of VQA tasks demonstrate that CROP significantly outperforms existing visual token pruning methods and achieves state-ofthe-art performance.

Recent advances in text-only "slow-thinking" reasoning have prompted efforts to transfer this capability to vision-language models (VLMs), for training visual reasoning models (VRMs).However, such transfer faces critical challenges: Effective "slow thinking" in VRMs requires visual reflection, the ability to check the reasoning process based on visual information.Through quantitative analysis, we observe that current VRMs exhibit limited visual reflection, as their attention to visual information diminishes rapidly with longer generated responses.To address this challenge, we propose a new VRM Reflection-V 1 , which enhances visual reflection based on reasoning data construction for cold-start and reward design for reinforcement learning (RL).Firstly, we construct vision-centered reasoning data by leveraging an agent that interacts between VLMs and reasoning LLMs, enabling cold-start learning of visual reflection patterns.Secondly, a visual attention based reward model is employed during RL to encourage reasoning based on visual information.Therefore, Reflection-V demonstrates significant improvements across multiple visual reasoning benchmarks.Furthermore, Reflection-V maintains a stronger and more consistent reliance on visual information during visual reasoning, indicating effective enhancement in visual reflection capabilities.

Retrieval-Augmented Generation (RAG) has demonstrated considerable effectiveness in open-domain question answering.However, when applied to heterogeneous documents, comprising both textual and tabular components, existing RAG approaches exhibit critical limitations.The prevailing practice of flattening tables and chunking strategies disrupts the intrinsic tabular structure, leads to information loss, and undermines the reasoning capabilities of LLMs in multi-hop, global queries.To address these challenges, we propose TableRAG, an SQL-based framework that unifies textual understanding and complex manipulations over tabular data.TableRAG iteratively operates in four steps: context-sensitive query decomposition, text retrieval, SQL programming and execution, and compositional intermediate answer generation.We also develop HeteQA, a novel benchmark designed to evaluate the multi-hop heterogeneous reasoning capabilities.Experimental results demonstrate that TableRAG consistently outperforms existing baselines on both public datasets and our HeteQA, establishing a new state-of-the-art for heterogeneous document question answering.

Recent advances have demonstrated that integrating reinforcement learning with rule-based rewards can significantly enhance the reasoning capabilities of large language models (LLMs), even without supervised fine-tuning (SFT). However, prevalent reinforcement learning algorithms such as GRPO and its variants like DAPO, suffer from a coarse granularity issue when computing the advantage. Specifically, they compute rollout-level advantages that assign identical values to every token within a sequence, failing to capture token-specific contributions. To address this limitation, we propose Key-token Advantage Estimation (KTAE)—a novel algorithm that estimates fine-grained, token-level advantages without introducing additional models. KTAE leverages the correctness of sampled rollouts and applies statistical analysis to quantify the importance of individual tokens within a sequence to the final outcome. This quantified token-level importance is then combined with the rollout-level advantage to obtain a more fine-grained token-level advantage estimation. Empirical results show that models trained with GRPO+KTAE and DAPO+KTAE outperform baseline methods across five mathematical reasoning benchmarks. Notably, they achieve higher accuracy with shorter responses and even surpass R1-Distill-Qwen-1.5B using the same base model.

Visual question answering (VQA) tasks, often performed by visual language model (VLM), face challenges with long-tail knowledge.Recent retrieval-augmented VQA (RA-VQA) systems address this by retrieving and integrating external knowledge sources.However, these systems still suffer from redundant visual information irrelevant to the question during retrieval.To address these issues, in this paper, we propose LLM-RA , a novel method leveraging the reasoning capability of a large language model (LLM) to identify key visual entities, thus minimizing the impact of irrelevant information in the query of retriever.Furthermore, key visual entities are independently encoded for multimodal joint retrieval, preventing cross-entity interference.Experimental results demonstrate that our method outperforms other strong RA-VQA systems.In two knowledge-intensive VQA benchmarks, our method achieves the new state-of-the-art performance among those with similar scale of parameters and even performs comparably to models with 1-2 orders larger parameters.

Falls are fatal for the elderly, and timely detection after falls is crucial. As a contactless device, the radar sensor can monitor users’ falls with the advantage of not revealing their privacy. Today, the use of radar for fall detection has made a significant progress. However, most current methods cannot be used in real complex scenes. They usually collect fewer types of actions, and the ratio of the number of nonfall samples to the number of fall samples is small, which is not consistent with the real-life scene. In addition, the classifiers are usually trained and tested on the same environments and the same people, which cannot be easily extended to new environments and new people. We designed a robust fall detection system based on the frequency-modulated continuous-wave (FMCW) radar to solve these issues. The system detects the moment of human movement and calculates the range–velocity map, range–horizontal angle map, and range–vertical angle map of the radar signals, and creates three neural networks for these three signal maps. The stacking method of ensemble learning is used to fuse the time–space–velocity features extracted by the three neural networks to identify falls. The method was trained and tested on a data set consisting of ten scenes, 21 subjects, 52 nonfall action types, and 12 fall action types. The results show that on the test set containing only new environments and new subjects, the recall is 0.983 and the precision is 0.975.

Mode-division multiplexing (MDM) using few-mode fibers (FMFs) appears as a promising technology to increase fiber capacity by a few orders of magnitude and sustain the traffic demand for the decades to come. The potential of MDM lies in its ability to exploit multiple modes within a single optical fiber strand. Since 2011 and the first promising demonstrations, impressive progress has been made. In this article, we will show how the most recent advances in design and fabrication have improved the performance of FMFs and MDM systems, and allowed to turn research demonstrations into practical deployments.