arXiv:2609.39142v1 Announce Type: new Abstract: Can structured text replace vision for diagram reasoning? A wrong answer after textualization can arise because the representation omits information th...
Can structured text replace vision for diagram reasoning? A wrong answer after textualization can arise because the representation omits information the question needs, or because the solver fails to...
ZoomCap is a method for fine‑grained captioning that treats missing visual evidence as recoverable information. It uses a draft caption to guide conditional knowledge retrieval, OCR‑aware reading, coverage‑gap recovery, and adaptive visual zooming, focusing deeper crops on complex regions and deriving spatial relations from grounded geometry. The accompanying FactCycle framework evaluates caption claims against the source image and a reconstruction, distinguishing fabrication from generator loss, and the approach shows strong performance gains in both general and medical vision‑language pre‑training.
arXiv:2604.03114v2 Announce Type: replace-cross Abstract: Vision-language models (VLMs) may need to forget visual concepts after deployment because of privacy, copyright, licensing, safety, or policy...
arXiv:2607. 25467v1 Announce Type: cross Abstract: Stateful multimodal assistants encode an image once but may answer questions about it many turns later.
arXiv:2607. 12364v1 Announce Type: cross Abstract: EEG-to-image evaluation should distinguish visual fidelity from recoverable meaning.
SCORE: Subject Coordinate Recovery for Label-Free Cross-Subject EEG-to-Image Retrieval proposes a new framework that aligns EEG signals from different subjects into a common image space without requiring labeled calibration data. By training on source subjects and estimating an orthogonal transformation at deployment, SCORE recovers target EEG coordinates and selects reliable EEG-image landmarks through hubness-corrected matching. The method achieves state‑of‑the‑art Top‑1/Top‑5 accuracy on two public benchmarks, outperforming existing baselines by significant margins.
arXiv:2606. 29788v1 Announce Type: new Abstract: When a multimodal AI agent is asked to forget a fact, current memory systems usually delete the text entry and report success.
The paper investigates how memory systems can answer a current query correctly yet fail to retain distinctions needed for later updates. Using a paired‑history audit, the authors evaluate 24 history pairs across six synthetic mechanisms and two model backends, achieving perfect reveal accuracy on DeepSeek and high accuracy on GLM. Record‑level audits reveal specific failures in structured reveal memories and frontier late‑reference adequacy, and the authors test a label‑equivariant repair that only partially restores correctness.
arXiv:2606.03401v2 Announce Type: replace Abstract: Scientific images function as critical evidence in research communication, yet their integrity faces unprecedented threats from AI-generated conten...
arXiv:2610.19611v1 Announce Type: new Abstract: Modern vision-language models can accept multi-image contexts, but aggregate accuracy rarely reveals where a multi-image reasoning pipeline fails. We s...
CVE‑SAI is a method for e‑commerce product attribute indexing that first infers a candidate attribute from the primary image and question, then uses counterfactual visual evidence to decide whether to admit that candidate into the index. It employs Focus‑Zone Distortion to create an attribute‑specific visual‑dependence proxy and Evidence‑Guided Attention Redistribution to refine scoring, ensuring that only visually supported and risk‑controlled attributes are indexed. Experiments on Amazon Berkeley Objects demonstrate improved attribute inference, evidence localization, and controlled retrieval performance with minimal unsafe exposure.