Multimodal Models - Weekly Roundup

Geometry-Change VLA (GC-VLA), which learns to predict multiview future-current geometry-change tokens from current observations and applies Geometry-Conditioned Residual Flow (GCRF), using a binary intervention router and a single bounded residual velocity policy learned from closed-loop feedback.

A hierarchical framework that couples high-level multimodal reasoning with low-level function-constrained execution for deployable precision medical robotics, and introduces a scalable multi-agent pipeline to generate skill-oriented chain-of-thought data for structured training.

A sharp width--magnitude tradeoff at fixed depth is established using one elementary bounded $1$-Lipschitz Dyadic--Triangular Activation and yields a continuum of statistically optimal choices, ranging from unit parameter radius to fixed network size.

CableVLA is presented, an end-to-end multimodal vision-language-action framework that converts simulation-privileged supervision into deployable cable-topology and tactile representations and cross-simulator and real-robot comparisons further examine zero-shot policy transfer under changes in dynamics and sensing.

Timbre drift in multi-turn expressive voice agents can be eliminated by replacing implicit context embeddings with explicitly supervised style instructions optimized via rejection sampling and context-aware DPO.

DSR is introduced, a causal post-selection operator that repairs destination support without retraining the host predictor or increasing the maintained set size, and shows that finite-set support allocation is a useful prediction-side control point when a fixed hypothesis set serves as the interface to downstream systems.

BAS-OPD is proposed, a budget-aware selective OPD framework that allocates teacher supervision under limited query budgets that achieves strong performance while substantially reducing teacher supervision costs, highlighting the effectiveness of selective OPD under constrained budgets.

OmniFysics-Nano-V2, a compact omni-modal model for physical-world perception and understanding, is presented and experiments show that the proposed data and training strategy improves physical-world understanding while preserving broad omni-modal competence.

A Unification Framework for Multimodal Data (UFMD) is introduced, which automatically processes tongue and pulse images into structured, clinically standard descriptions, integrating multi-source diagnostic information into a unified digital record of I-AOI-P process.

Missing sensory streams typically break audiovisual tracking, but routing audio, face, and mouth inputs through modality-agnostic query attention shatters previous active speaker detection records across five benchmarks by up to 5.1 mAP.

SASS (Stage-Adaptive Sample Selection), a stage-adaptive data-selection framework for pool-based active learning in long-tailed dense prediction, shows that annotation-efficient learning depends not only on which samples are selected, but also on how the annotation budget is distributed across categories and when model-derived scores begin to guide selection.

PartLLM is introduced, a unified multimodal model that formulates 3D part segmentation as autoregressive semantic decomposition that supports text-guided part segmentation, interactive segmentation, and full-shape semantic decomposition with controllable granularity within a single model.

The results establish explicit-implicit matching as an effective and efficient mechanism for spatio-temporal instance interaction and Integrating temporal MatchFusion into SparseDrive further improves perception within an E2E framework without additional supervision.

Findings show that configuration changes can affect how models use information they can still read, and attention interventions in LLaVA-NeXT suggest that configuration changes can weaken the use of readable information during answering.

STD is proposed, a hierarchical token pruning framework that adapts token selection mechanisms to the functional role of each network stage, and introduces a Stability-Adaptive Trigger in deep layers to execute pruning only during semantically stable phases.

SALI-FP is introduced, an evidence-gated multimodal pipeline that converts a plan into reviewable semantic maps, objects, vectors, and relation records while constraining local revisions by image evidence.

A surgical instrument-state logger that clips onto a conventional instrument without modifying the part that enters the patient and fuses a six-axis IMU and a time-of-flight rangefinder with a markerless camera in an error-state Kalman filter under the remote center of motion (RCM) of the trocar is presented.

This work finds that point clouds are substantially more reliable than RGB under zero-shot category shift, and designs a Multi-view Information Decoupling module to suppress heterogeneous information from multi-view point cloud projections and enhance point cloud feature quality.

Tracking subtle post-radiotherapy tissue evolution across all four standard mammographic views yields substantially richer prognostic signals when temporal differences are explicitly isolated prior to clinical fusion.

A training-free agentic framework built around specialist detectors, per-detector triage, and conflict-aware evidence arbitration is studied, indicating that the main challenge of open-world image forensics is not detecting manipulations, but calibrating trust in specialized forensic tools and arbitrating conflicting evidence.

Qwen-Audio-3.1-Realtime brings these requirements together through Think, Act, and Speak and Coordinate, evaluating audio reasoning, multilingual understanding, tool use, conversational behavior, full-duplex interaction, and safety.

Document Retrieval-Aware Chunking (D-RAC), an extension of the Web Retrieval-Aware Chunking (W-RAC) framework to arbitrary document formats, is presented, preserving W-RAC's cost, determinism, and observability benefits while unlocking every renderable format as a first-class input.

VideoGen-Agent is presented, a multimodal agent trained through multitask agentic reinforcement learning to use external tools for video generation to support learning tool use across video-generation tasks and show that the trained agent can benefit from subsequent advances in generation tools.