Vision-language models, speech and cross-modal systems that read, look and listen in the same forward pass.
arXiv:2609.16233v1 Announce Type: cross Abstract: Vision-language models excel at 2D image understanding but remain limited in 3D spatial reasoning. Progress is hindered by limitations in current ben...
arXiv:2603.12264v2 Announce Type: replace Abstract: Unified multimodal models target joint understanding, reasoning, and generation, but current image editing benchmarks are largely confined to natur...
arXiv:2609.17499v1 Announce Type: cross Abstract: Uncertainty estimation for Vision-Language-Navigation (VLN) models is a critical task since it can help identify ambiguous and unreliable predictions...
arXiv:2609.16284v1 Announce Type: cross Abstract: Query-conditioned vision--language models enable fine-grained interpretation by revealing how visual evidence changes with textual queries. However,...
arXiv:2609.17475v1 Announce Type: new Abstract: Capable open-weight models make local coding and reasoning attractive, but their context and execution state strain laptop memory. We present JustFit,...
arXiv:2609.17421v1 Announce Type: new Abstract: Audio-Visual Navigation (AVN) requires an agent to localize and navigate toward a continuously vocalizing target relying solely on visual observations...
arXiv:2609.17180v1 Announce Type: new Abstract: Multimodal counselor response generation (MCRG) aims to generate an appropriate counselor response from multimodal dialogue histories. Progress is limi...
arXiv:2609.17064v1 Announce Type: new Abstract: Assistive autonomous systems must anticipate human goals before an observed behavior is complete. This article formulates anticipation as goal inferenc...
arXiv:2609.16204v1 Announce Type: new Abstract: Safety guardrails in open-weight language models can be readily bypassed using Refusal Feature Ablation (RFA), a technique that identifies and projects...
MechReason is a new benchmark for multi-image, multi-hop reasoning in mechanical engineering, featuring 12,000 question-answer pairs with explicit reasoning chains and 21,000 visual items across nine evidence types. It covers eight task types—explanation, prediction, design, and diagnosis—within four reasoning dimensions, and is constructed through a four-stage pipeline that extracts engineering claims, generates verification-masked questions, and validates multimodal quality. Even state‑of‑the‑art models score only 62.89% on this dataset, highlighting its difficulty.
The paper introduces VPRef, the first cross‑domain benchmark for Referring Remote Sensing Image Segmentation, containing 46,972 language‑image‑annotation triplets with a three‑tier linguistic hierarchy. It proposes a parameter‑efficient adaptation method based on the Segment Anything Model and Low‑Rank Adaptation, using pseudo‑label self‑training for visual drift and random multi‑granularity prompt mixing for textual drift. Experiments show the approach improves cross‑domain segmentation while altering only 1.08 % of the base model’s parameters, offering a strong baseline for future research.
The paper introduces ReImaGin, a method that uses image generation models as a flexible visual reasoning tool for multimodal large language models. Unlike traditional fixed-function vision tools, ReImaGin accepts natural language commands and can perform open-ended visual operations such as removing occlusions or creating floorplans from multiple views. Experiments on six diverse visual reasoning tasks show that ReImaGin outperforms both text-only reasoning and specialist vision-tool baselines, achieving up to a 25% improvement.
EgoPathBench is a new dataset and benchmark that tests zero‑shot egocentric waypoint decision‑making in vision‑language models. Each task presents an egocentric RGB image, a natural‑language goal, and numbered visible waypoints, and models must return traversable candidates or an ordered route. The benchmark evaluates candidate feasibility, edge legality, and goal arrival under point‑agent or embodied geometry, covering 31,852 training, 1,345 validation, and 1,111 benchmark questions. "whyItMatters":"The benchmark reveals that current VLMs perform poorly on integrated navigation tasks, highlighting a gap in spatial intelligence that can be addressed by fine‑tuning with the released training data."
arXiv:2609.16368v1 Announce Type: cross Abstract: Vision-language models (VLMs) can generate routes directly from aerial imagery for off-road navigation, but their predictions provide no indication o...
SAVOR is a training framework for multimodal large language models that adds token and answer confidence to the output schema, optimises a Group Relative Policy Optimisation objective to penalise calibration error and poor abstention, and uses the learned confidence at inference to revisit visual evidence only when uncertain. Experiments on POPE, HallusionBench, AMBER, and MMHal-Bench with InternVL3-8B and Qwen3-VL-8B backbones show that SAVOR reduces hallucination while maintaining general capability on MME and MMBench, achieving lower Expected Calibration Error than DPO and decoding baselines.
arXiv:2609.16646v1 Announce Type: new Abstract: When strong multimodal models are widely available, progress requires new scientific methodologies beyond benchmark scores---using models as instrument...
Hyper-RED introduces a scalable image-to-event pretraining framework that transfers high‑order semantic structures via hypergraphs, avoiding rigid pixel‑wise alignment. By constructing image, event, and cross‑modal hypergraphs and applying a hypergraph relational distillation loss, the method preserves local relational consistency and event‑specific characteristics while inheriting image‑derived semantic organization. Experiments across five event datasets show consistent scaling from ViT‑S to ViT‑L and state‑of‑the‑art performance.
NeuroSymbEAD is a large‑scale neuro‑symbolic caption dataset that builds an ego‑centric knowledge graph of static and dynamic objects on the KITTI‑360 dataset, annotating classes, categories, heading directions, orientations, and distances from the ego‑vehicle. The dataset generates multilevel textual captions that serve as a lightweight representation of an ego‑centric scene map, enabling outdoor scene‑map reconstruction, visual recognition, and object grounding. Baselines for driving common sense and traffic/scene understanding are established, and the dataset is benchmarked using pre‑trained grounding and learned auto‑regressive captioning networks to support vision‑language and foundation models for traffic‑scene explanation, 3D reasoning, and interpretable autonomous‑driving perception.
The paper introduces the Multi-modal Knowledge Preserving Adapter (MKP-Adapter), an adapter-only approach that enables backward compatible training for multi-modal large language models without updating the backbone. It employs a multi-level preservation loss to maintain embedding geometry and a focal re-weighting strategy to focus on difficult samples. Experiments show strong backward compatibility across image, text, visual document, and video retrieval tasks with minimal latency overhead.
Video-HolmesV2 is a new benchmark that tests multimodal large language models on their ability to reason with spatio‑temporal audio‑visual evidence in long videos. It requires models to justify answers with precise evidence, uses a multi‑model cross‑verification pipeline and a spatio‑temporal evidence‑aware metric, and introduces an audio‑text guided token compression framework to reduce long‑context noise. In evaluations, even strong proprietary models score below 60% while the proposed approach outperforms comparable open‑source omni‑models.