The paper introduces CLEAR, a latent Flow Matching framework designed to generate visual counterfactual explanations (VCEs) that alter a classifier’s prediction while preserving realistic and decision‑irrelevant content. CLEAR addresses two main issues in latent‑space VCEs: opposing update directions between latent generation and classifier guidance, and unintended changes from VAE reconstruction. By decomposing classifier guidance into cooperative and conflicting components and selectively attenuating the conflicting part, CLEAR achieves counterfactual quality comparable to image‑space methods while being 2.5× faster.
arXiv:2609.37638v1 Announce Type: new Abstract: Current explanation methods for contrastive vision--language models such as CLIP mainly identify important regions without showing how to change the in...
The paper introduces a classifier‑free method for generating visual counterfactual explanations (VCEs) using Contrastive Analysis (CA). By separating generative factors common to two datasets from those specific to each class, the approach swaps only the salient factors to produce counterfactual images, thereby avoiding reliance on classifier decision boundaries. Leveraging StyleGAN2’s high‑quality synthesis and a feature‑space latent representation, the method supports multiple salient factors per dataset and achieves superior counterfactual quality on three medical imaging datasets.
arXiv:2607. 22544v1 Announce Type: new Abstract: Visual counterfactual explanations aim to answer "what minimal change to this image would flip the model's prediction?
DVLA-RL++ introduces Dual-Level Vision‑Language Alignment with Reinforcement Learning Gating for few‑shot learning. It adds complementary semantic purification (CSP) to separate intrinsic from nuisance textual descriptions and counterfactual reinforcement‑learning gating (CRG) to learn layer‑wise fusion strengths. Experiments on standard, fine‑grained, and cross‑domain benchmarks demonstrate state‑of‑the‑art accuracy, improving on DVLA‑RL by an average of 1.4%.
arXiv:2610.01942v1 Announce Type: new Abstract: Predicting the future evolution of a scene is a fundamental capability for world modeling. Recent work has shown that operating in the feature space of...
Self-improvement for multimodal large language models (MLLMs) is typically driven by reward-based methods that provide only coarse scalar feedback. Distillation offers a richer alternative through dense token-level supervision, but in the visual domain it usually depends on privileged context constructed using external annotations and tools, or stronger models.
The paper introduces visual adaptations of counterfactual tests—vCT and vCCT—to evaluate whether chain-of-thought explanations in vision‑language models faithfully reflect the visual evidence driving predictions. Using these tests, the authors benchmark eight open‑source VLMs on two datasets and find that CoTs often fail to track visual evidence, sometimes omitting removed objects or mentioning them inconsistently. They also release two new datasets, Counter‑SNLI‑VE and Counter‑A‑OKVQA, consisting of image pairs that differ by a single object to facilitate further research.
arXiv:2609.37165v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models remain brittle under visual distribution shifts, often relying on spurious correlations tied to domain-specific f...
The paper introduces Future‑L1, an interleaved latent visual reasoning framework that alternates between language tokens and continuous latent visual spans during autoregressive decoding for video event prediction. By training on a curated dataset (Future‑L1‑50K) and optimizing latent trajectories with a latent‑aware RL objective (LA‑DAPO), Future‑L1 achieves state‑of‑the‑art results on FutureBench and TwiFF‑Bench, outperforming prior models by significant margins. The work demonstrates that preserving intermediate visual semantics in latent space, rather than verbalizing every reasoning step, leads to more accurate and visually grounded future predictions.
Explainable Deepfake Detection with Feature-robust Augmentation and Evidence-grounded Explanation Optimization proposes a new framework that improves deepfake detection and interpretability. The approach introduces Feature-robust Augmentation—diversified degradation-aware strategies combined with supervised contrastive learning and a mean-teacher architecture—to maintain accuracy on low-quality images. For explanations, it employs evidence-grounded preference optimization, guiding the model to focus on genuine manipulation traces by learning from chosen-rejected explanation pairs that omit evidence or inject irrelevant details. The method achieved first place in the ACM Multimedia 2026 Explainable Deepfake Detection Challenge and is publicly available on GitHub.
CounterVid introduces a scalable counterfactual video generation framework that creates videos differing only in actions or temporal structure while keeping scene context intact. The approach uses multimodal LLMs for action proposals and diffusion models for editing, producing a synthetic dataset of ~26k preference pairs for action recognition and sequence ordering. With the MixDPO optimization method, the authors demonstrate significant improvements in action recognition and temporal ordering on Qwen2.5‑VL and InternVL3 backbones, while maintaining overall video understanding.
arXiv:2607. 16311v1 Announce Type: cross Abstract: Vision-language models (VLMs) often answer visual questions using learned language and category priors rather than grounding their predictions in the image itself.