home.social

#anomaly-detection — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #anomaly-detection, aggregated by home.social.

fetched live
  1. Anomaly detection in fintech fraud systems identifies transactions that deviate significantly from an established behavioral baseline for a specific user or account.

    #Fintech #AnomalyDetection

  2. Stop using "High/Medium/Low" labels for boardroom risk. It’s an insult to the math.

    ​The HFY Coefficient is a deterministic 0-100 scalar. We feed Shiki’s Latent Vectors into an Isolation Forest (iForest) engine and run 100,000-iteration "Kill-Shot" simulations.

    ​HFY = np.clip((0.5 - iForest.decision_function) * 100, 0, 100)

    ​It measures the exact mathematical distance between a vendor's reality and structural collapse. 📉

    ​#DataScience #AnomalyDetection #RiskEngineering #Math

  3. You already know that you can visualize your metrics from in Dashboard's Discover Metrics experience (if not, check the comments).

    But what if we could add some sauce to detect anomalies and extrapolate forecasts?

    Check out the new RFC for time series and in @OpenSearchProject and chime in with your feedback.
    github.com/opensearch-project/


    @Prometheus

  4. Many industries hope to benefit from , but they know little about where it works & where it’s unreliable. is a good field to test AI, with complex factors, interacting settings, & unpredictable conditions, says our author Andy Oram.

    Read more in this article: lpi.org/711y

  5. Anomaly Detection Analysis with Python
    Find unusual transactions without labels, using a baseline + Isolation Forest + practical verification.
    This post shows a clean workflow: define “unusual” with a baseline, train Isolation Forest, validate with simple sanity checks, and reduce false alarms with practical rules.

    🔗 medium.com/towards-artificial-

    #Python #DataScience #AnomalyDetection #MachineLearning #Fraud

    @chartrdaily @programming @pythonclcoding @theartificialintelligence @medium

  6. Nice story about #AI assisting a rescue mission in the alps: bbc.com/future/article/2026010

    Well, in this case it was slightly too late, it is an interesting use case for AI nonetheless. Note that this is not #LLMs obviously, but some kind of #AnomalyDetection for #ComputerVision.

  7. 🚨 New CRAN Task View: Anomaly Detection

    By Priyanga Dilini Talagala @pridiltal , Rob J. Hyndman @robjhyndman Gaetano Romano

    URL: CRAN.R-project.org/view=Anomal

  8. Most security systems are reactive, designed to catch a fire after it has already started. Our conceptual architectural blueprint includes a proactive, Context-Aware Anomaly Detection System that learns "normal" behavior and flags suspicious intent-not just malicious IP addresses. This is the difference between a clumsy shield and an intelligence-driven defense.
    #DataSecurity #AnomalyDetection #Al #MachineLearning #BehavioralAnalytics
    #ProactiveSecurity #Strategiclntelligence #ShaolinDataScience

  9. Institute for AI @UniStuttgartAI@bawü.social ·

    MultiADS: Defect-aware Supervision for Multi-type Anomaly Detection and Segmentation in Zero-Shot Learning

    In manufacturing, quality control remains a critical yet complex task, especially when multiple defect types are involved. MultiADS introduces a system capable of detecting and segmenting a wide range of anomalies (e.g., scratches, bends, holes), even in zero-shot settings.

    By combining visual analysis with descriptive textual input and using a curated Knowledge Base for Anomalies, MultiADS generalizes to unseen defect types without requiring prior visual examples and consistently outperforms state-of-the-art models across several benchmarks, offering a robust and scalable solution for industrial inspection tasks.

    Sadikaj, Y., Zhou, H., Halilaj, L., Schmid, S., Staab, S., & Plant, C. MultiADS: Defect-aware Supervision for Multi-type Anomaly Detection and Segmentation in Zero-Shot Learning. International Conference on Computer Vision, ICCV 2025, Hawai, Oct 19-23, 2025, #ICCV2025. arxiv.org/abs/2504.06740.

    #AI #AIResearch #ComputerVision #AnomalyDetection #ZeroShot

  10. 🚨🚂 Welcome aboard the 🚀 #AppSignal 🛤️ express, where buzzwords like "Solid Queue" sound like a hipster brunch choice and "Anomaly Detection" is your morning coffee spilling! ☕ Who knew Ruby on Rails needed more rails and less ruby? 🤷‍♂️
    blog.appsignal.com/2025/05/07/ #SolidQueue #AnomalyDetection #RubyOnRails #TechTrends #HackerNews #ngated

  11. Shared Nearest Neighbors (SNN) — A distance metric that can improve prediction, clustering, and outlier detection in datasets with many dimensions and with varying densities. Read more from W Brett Kennedy now!

    #Clustering #AnomalyDetection

    towardsdatascience.com/shared-