arXiv:2606. 30196v1 Announce Type: cross Abstract: This paper offers an in-depth analysis of non-sequential multimodal sentence-level embeddings, with a particular focus on the SONAR model.
arXiv:2507.12295v2 Announce Type: replace-cross Abstract: Text anomaly detection is a critical task in natural language processing (NLP), with applications spanning fraud detection, misinformation id...
arXiv:2608. 13652v1 Announce Type: new Abstract: Generic event-level anomaly detection for collider physics has two recurring problems: anomaly scores are hard to interpret, and they correlate strongly with energy scale and object multiplicity.
arXiv:2609.08200v1 Announce Type: new Abstract: Token-level text anomaly detection, as an emerging trend of text anomaly detection, moves beyond coarse-grained document-level detection by localizing...
arXiv:2606. 01992v1 Announce Type: cross Abstract: Industrial anomaly detection has historically been a unimodal task.
Recent multimodal LLM (MLLM)-based anomaly detection (AD) methods can generate textual explanations that help interpret their predictions. However, reliably and efficiently evaluating these explanatio...
arXiv:2606. 30675v1 Announce Type: cross Abstract: Early detection of dementia through speech analysis offers a non-invasive screening alternative, but capturing both acoustic and linguistic biomarkers remains challenging.
The paper introduces a statistical feature augmentation technique that encodes behavioral interaction statistics into the input space for dynamic graph anomaly detection. Experiments on Reddit, Wikipedia, and MOOC datasets across seven models—both continuous-time and discrete-time—show that this augmentation consistently improves detection performance compared to models trained on original embeddings. The enriched input also facilitates fine-grained post-hoc analysis of behavioral importance, linking classical network analysis with deep learning.
arXiv:2610.27565v1 Announce Type: cross Abstract: Recent multimodal LLM (MLLM)-based anomaly detection (AD) methods can generate textual explanations that help interpret their predictions. However, r...
arXiv:2607. 12290v1 Announce Type: cross Abstract: Audio-language embedding models such as CLAP are widely evaluated on matching present sound events, but rarely on negation.
Probe‑VAD introduces an ordinal binary‑probing framework that leverages frozen vision‑language models for training‑free video anomaly detection. By querying ten ordered severity thresholds and extracting YES/NO continuation likelihoods, it builds a cumulative severity profile that is converted into a continuous anomaly score with isotonic projection for ordinal consistency. Experiments on public benchmarks show that this simple interface yields superior performance at low computational cost, avoiding the limitations of caption‑based compression or restricted numerical scoring.
Vision foundation models have enabled strong training-free anomaly detection (AD). However, most existing approaches rely primarily on independent local patch features, leaving the global contextual information encoded by Vision Transformers (ViTs) underexploited.
arXiv:2508. 00955v3 Announce Type: replace-cross Abstract: Adapting generative Multimodal Large Language Models (MLLMs) into universal embedding models typically demands resource-intensive contrastive pre-training, while traditional hard negative mining methods suffer from severe false negative contamination.