Tourism and travel planning increasingly rely on digital assistance, yet existing multimodal AI systems often lack specialized knowledge and contextual understanding of urban environments. We present TraveLLaMA, a specialized multimodal language model designed for comprehensive travel assistance. Our work addresses the fundamental challenge of developing practical AI travel assistants through three key contributions: (1) TravelQA, a novel dataset of 265k question-answer pairs combining 160k text QA from authentic travel sources, 100k vision-language QA featuring maps and location imagery, and 5k expert-annotated Chain-of-Thought reasoning examples; (2) Travel-CoT, a structured reasoning framework that decomposes travel queries into spatial, temporal, and practical dimensions, improving answer accuracy by 10.8% while providing interpretable decision paths; and (3) an interactive agent system validated through extensive user studies. Through fine-tuning experiments on state-of-the-art vision-language models (LLaVA, Qwen-VL, Shikra), we achieve 6.2-9.4% base improvements, further enhanced by Travel-CoT reasoning. Our model demonstrates superior capabilities in contextual travel recommendations, map interpretation, and scene understanding while providing practical information such as operating hours and cultural insights. User studies with 500 participants show TraveLLaMA achieves a System Usability Scale score of 82.5, significantly outperforming general-purpose models and establishing new standards for multimodal travel assistance systems.

Xichen Zhang, Ziyi He, Yinghao Zhu, Sitong Wu, Shaozuo Yu, Meng Chu, Wenhu Zhang, Haoru Tan, Jiaya Jia. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

We present VRBench, the first long narrative video benchmark crafted for evaluating large models' multi-step reasoning capabilities, addressing limitations in existing evaluations that overlook temporal reasoning and procedural validity. It comprises 960 long videos (with an average duration of 1.6 hours), along with 8,243 human-labeled multi-step question-answering pairs and 25,106 reasoning steps with timestamps. These videos are curated via a multi-stage filtering process including expert inter-rater reviewing to prioritize plot coherence. We develop a human-AI collaborative framework that generates coherent reasoning chains, each requiring multiple temporally grounded steps, spanning seven types (e.g., event attribution, implicit inference). VRBench designs a multi-phase evaluation pipeline that assesses models at both the outcome and process levels. Apart from the MCQs for the final results, we propose a progress-level LLM-guided scoring metric to evaluate the quality of the reasoning chain from multiple dimensions comprehensively. Through extensive evaluations of 12 LLMs and 19 VLMs on VRBench, we undertake a thorough analysis and provide valuable insights that advance the field of multi-step reasoning.

Long-form video understanding (LVU) addresses the challenge of answering complex questions over extended video length, where informative cues are sparse and easily overwhelmed by redundant content. To tackle this, it requires selecting a small set of question-relevant keyframes and reasoning over long-range, temporally dispersed visual evidence. However, current methods typically extract frame-level features with limited temporal context and store them in sequential memory structures. As a result, they struggle to capture the evolving relations among entities and fail to maintain identity consistency when entities temporarily leave and later reappear in the video. These limitations prevent accurate keyframe localization and coherent reasoning.

The combination of teleoperating robots and the Internet of Things (IoT) could be employed in many areas, including remote nursing and semi-mechanical control. However, it is known that subjects can quickly encounter physical and mental fatigue, which can potentially lower accuracy in teleoperation. To address this issue, this article presents a closed-loop teleoperation system based on multisensory fusion, visual, and haptic feedback within the IoT framework. Various sensors for electromyography, inertial measurement unit, and mechanical hand control system are deployed to obtain body signals from participants, subsequently processed by artificial intelligence methods. Resistive sensors are installed at the robotic hand side to learn the contact force between the robot hand and the grasped object. To allow the user to understand the contact force levels, a haptic interface is equipped to provide three-level mechanical vibrations. To help users identify and trace different objects, a region convolution neural network for detecting 100 different categories of things is constructed. In addition, the real-time control (delay$\mathbf {< } 110$ms) is offered by a tri-layered IoT architecture. The feasibility of the proposed technique is validated by a grasping task carried out by 20 volunteers. During experiments, it is observed that merely 10.8 s is averagely needed for participants in finishing the task, with a high average success rate of 97%. Besides, NASA-TLX questionnaire and maximum voluntary contraction test report that users suffer light mental and physical fatigue when the proposed technique is used.