TimeSeek pretrains one model for multi-task time series analytics
After large-scale pretraining, authors report state-of-the-art forecasting, anomaly detection, and classification in few-shot and zero-shot settings.
Stories per day · 30 days
2 RSS
Relatedtime-series3foundation-models2industrial-control25g1adversarial-disturbances1ai-agents1asvspoof1audio1aviation1behavioral-monitoring1benchmark1beth-dataset1botnet1capacity-planning1cyber-physical1deepfake-detection1diffusion-models1dinov21
45 stories, newest first · one card per story
Stories tagged #anomaly-detection
Survey: security teams plan most monitoring investment for AI agents
Nearly half of respondents said their organizations cannot effectively baseline normal AI agent behavior.
1
Vendi Anomaly Score flags outliers via dataset diversity
The score needs no hyperparameter tuning, scales linearly, and led 148 embedding benchmarks plus ImageNet tests.
Static ramp model leaves detectable residuals on Siemens PROFINET drives
PASAD and CUSUM flag the omitted overshoot residual, and evaluated N4SID models did not improve stealth.
Attack-free WavLM detector ranks audio deepfakes but fails under shift
The best variant scores 8.45% EER on ASVspoof 2019 LA, then 22.12% EER on the 2021 DF shift.
Fast-Tier Models Miss Held-Out Web Attack Families
ModernBERT classifies 37-60% of held-out path traversal as benign within 5% budget; specialized models escalate 97.6% but may include public families in training.
Gfm4Mtsad adapts graph foundation models for MTSAD
Gfm4Mtsad bridges pretrained GFMs with time series data using SICA, MAPS and TGEC modules for superior anomaly detection.
Dynamic Graph Models Detect Anomalies in 5G Networks
Proposes dynamic graph formalism for 5G anomaly detection capturing packet-level content and contexts.
Stacked hardware controller enhances Edge AI security
A stacked hardware security controller enables low-overhead anomaly detection in Edge AI systems using performance counters.
PairWatch spots adversarial satellite rendezvous in public catalogs
Only PairWatch kept one threshold that separated real rendezvous from operational negatives in every bootstrap resample.
Industrial detectors that look deployable in simulation fail on a real plant
At a fixed false-alarm budget, sequential valve attacks were flagged on at most 0.17 of perturbed samples.
Pneuma releases aircraft time-series data for predictive maintenance
The public set has over 920 million steps, four aircraft types, and nine verified fault scenarios.
Distributed log parsers tested across 6,720 federated configurations
Five parsers plus aggregation methods were compared so clients can agree on event templates without ground-truth logs.
Score ranks stealthy source manipulation of anomaly detectors
On a 17-channel telemetry benchmark, a detector blind to 64.7% of audited task and subset pairs still caught 82.8% of random single-channel corruption.
CLOSURE detects time-series anomalies via consistency and recoverability
A diffusion reconstructor and a relation-closure operator are trained separately and calibrated only on normal data.
SWAPE jointly learns parameters and sparse sample weights
Alternating majorize-minimize updates cover convex and difference-of-convex losses while down-weighting anomalous points.
Flownomaly recovers normal features for few-shot anomaly detection
In the 1-shot setting, image-level AUROC reached 97.8% on MVTec-AD and 95.3% on VisA.
WADE flags anomalous WebAssembly behavior with per-app LSTM models
On 67 Wasm apps and ten vulnerability classes, the authors report zero false negatives and a 0.42% false positive rate.
Hybrid fuzzy ensemble reports 0.9951 AUROC on BETH botnet detection
A TSK classifier plus Fuzzy C-Means and a Min Distance metric beats Isolation Forest by 17.1% on the BETH test set.
96-day study measures SRT streaming behavior and UDP flood effects
Controlled 500-pps UDP floods raised CPU 176–806% while SRT metrics moved only about 13%.
Tabular anomaly detection fuses foundation-model semantics with domain features
Authors report the method beats competitors when training uses only eight labeled anomalies.
Local Diagnostics of Learned Event Representations for Anomaly Detection at the HL-LHC
Supervised embeddings improve HL-LHC anomaly detection on 750 million simulated events and add local diagnostics.
Pareto-optimal quantum kernel selection for unsupervised anomaly detection on real malware beaconing data
Pareto quantum-kernel selection on real malware beaconing slightly beats a classical one-class SVM baseline.
GCTAuto-encoder: A Cross modal Framework for Security Flaw Detection in IoT Networks
GCT auto-encoder reports 0.908 accuracy detecting IoT network intrusions from edge data.