VLMs show a resolution illusion on UHR Earth observation imagery where higher resolution does not improve micro-target perception; UHR-Micro benchmark and MAP-Agent address this via evidence-centered active inspection.
hub
Learning transferable visual models from natural language supervision
34 Pith papers cite this work. Polarity classification is still indexing.
34
Pith papers citing it
hub tools
citation-role summary
background 2
citation-polarity summary
co-cited works
years
2026 34roles
background 2polarities
background 2representative citing papers
SAEParate disentangles sparse representations in diffusion models via contrastive clustering and nonlinear encoding to enable more precise concept unlearning with reduced side effects.
PoseBridge recovers semantic information lost during skeletonization by extracting pose-anchored cues from human pose estimation and transferring them via skeleton-conditioned bridging and semantic prototype adaptation, yielding 13.3-17.4 point gains on the Kinetics PURLS benchmark.
PhyGround is a new benchmark with curated prompts, a 13-law taxonomy, large-scale human annotations, and an open physics-specialized VLM judge for evaluating physical reasoning in generative video models.
GridProbe uses posterior probing on a KxK frame grid to adaptively select question-relevant frames, delivering up to 3.36x TFLOPs reduction with accuracy within 1.6 pp of the full-frame baseline on Video-MME-v2.
LoopVLA adds recurrent refinement and learned sufficiency estimation to VLA models, cutting parameters 45% and raising throughput 1.7x while matching baseline task success on LIBERO and VLA-Arena.
IPAD-CLIP adapts CLIP via artifact-aware text embeddings to detect multi-class local perceptual artifacts, backed by a new dataset of 3520 images with pixel-level masks.
SphereVAD performs training-free video anomaly detection by recasting anomaly discrimination as von Mises-Fisher likelihood-ratio geodesic inference on the unit hypersphere using intermediate MLLM features, with Frechet mean centering, holistic scene attention, and spherical geodesic pulling.
Excess risk decomposes into independent alignment (trace of inverse average Hessian times gradient covariance) and curvature terms, so both flatness and gradient alignment are required; SAGE achieves this and sets new SOTA on DomainBed.
PIQL integrates train-time-only privileged information into tabular foundation models via new constructions and a reconstruction architecture to achieve faster convergence and better generalization.
Optimal INR freeze depth matches highest weight stable rank layer; SAEs reveal SIREN atoms are localized while FFMLP atoms trace cohort contours with causal impact on PSNR.
S2M extracts structured text quadruples from change masks to provide noise-free multimodal supervision, achieving 17.80% Sek and 66.14% F_scd on the new Gaza-Change-v2 dataset and outperforming LLM-based multimodal methods.
TrajGANR learns continuous neural representations of trajectories to enable fine-grained alignment with street-view images and locations in a joint multimodal self-supervised objective, outperforming prior geospatial MSSL methods on urban mobility and road tasks.
SyncDPO improves temporal synchronization in video-audio joint generation using DPO with efficient on-the-fly negative sample construction and curriculum learning.
LDDR proposes a linear DPP-based dynamic-resolution frame sampler that achieves 3x speedup and up to 2.5-point gains on video MLLM benchmarks by selecting non-redundant frames and allocating tokens accordingly.
SpatialForge synthesizes 10 million spatial QA pairs from in-the-wild 2D images to train VLMs for better depth ordering, layout, and viewpoint-dependent reasoning.
Super-Linear Advantage Shaping (SLAS) introduces a non-linear geometric policy update for RL post-training of text-to-image models that reshapes the local policy space via advantage-dependent Fisher-Rao weighting to reduce reward hacking and improve performance over GRPO baselines.
CLEF, a long-context EEG foundation model using 3D multitaper spectrograms and contrastive alignment with reports and EHR, beats prior models on 229 of 234 clinical tasks and raises mean AUROC from 0.65 to 0.74.
ROMA improves MLLM robustness to seen and unseen visual corruptions by +2.3-2.4% over GRPO on seven reasoning benchmarks while matching clean accuracy.
LLaVA-UHD v4 reduces visual-encoding FLOPs by 55.8% for high-resolution images in MLLMs via slice-based encoding plus intra-ViT early compression while matching or exceeding baseline performance on document, OCR, and VQA benchmarks.
MLS-Bench shows that current AI agents fall short of reliably inventing generalizable ML methods, with engineering tuning easier than genuine invention.
DeCIR decouples endpoint alignment from semantic transition alignment in projection-based ZS-CIR via paired edit tuples, separate low-rank adapters, and LRDM merging, yielding consistent gains on CIRR, CIRCO, FashionIQ, and GeneCIS without added inference cost.
DVD treats voxel occupancy as a discrete variable in a diffusion framework to generate, assess, and edit sparse 3D voxels without continuous thresholding.
Modality representations share dominant semantic geometry but have an anisotropic residual gap; AnisoAlign corrects source representations boundedly using target geometry for unpaired alignment.
citing papers explorer
-
LithoBench: Benchmarking Large Multimodal Models for Remote-Sensing Lithology Interpretation
LithoBench is a new multi-level benchmark showing that existing large multimodal models have substantial limitations in geological semantic understanding for remote sensing lithology interpretation.