proomt

Search

Search posts, papers, and topics

All posts

Hugging Face Daily PapersChengxian Hu, Zhiming Ma, Mingjun Pan1 min readpaperadvanced

FRAUDSkill: Structured Frozen-Weight Skill Optimization for Audio Anti-Fraud Detection

Read on arxiv.org

Summary

FRAUDSkill is a framework that keeps a pretrained audio‑language model frozen and learns an external layer of skill programs, routing policies, and decision rules to meet a structured anti‑fraud detection protocol. On the TeleAntiFraud benchmark it reaches 73.5% Macro‑F1 (≈32% improvement) while cutting invalid predictions to 1.94%.

  • FRAUDSkill adds an external skill layer to a frozen audio‑language model, avoiding any model weight updates.
  • Route‑specific policies and decision rules enforce the required multi‑step fraud detection protocol.
  • Validation‑guided multi‑path inference reduces invalid outputs to 1.94%.
  • Achieves 73.5% Macro‑F1 on TeleAntiFraud, a 31.96% improvement over the frozen baseline.

Security engineers and ML practitioners building audio‑based fraud detection systems should care because it offers a plug‑in adaptation path that improves accuracy and compliance without costly model retraining.

7/10
llmsecurityml infraaudiospeech

Related reading

  1. TeleAntiFraud 2.0: A Refreshable, Profile-Grounded, and Audio-Based Benchmark for Telecom Fraud Detection

    The paper introduces TeleAntiFraud 2.0, a monthly refreshed, frozen audio benchmark of 900 Chinese telecom calls (600 fraud, 300 near‑domain non‑fraud) built with a Mixed‑Tree generation pipeline. Experiments show models that score perfectly on unrelated negatives fall to ~0.66 Macro‑F1 on near‑domain cases, exposing shortcut learning and prediction collapse.

    Hugging Face Daily Papersarxiv.org1 minpaper

  2. Grounded Skill Synthesis from Code at Scale for Agentic Intelligence

    Code2Skill is an automated pipeline that mines popular GitHub repositories to extract verifiable, implementation‑anchored procedural “skills”. It builds a bank of ~1 M skill records (atomic ops, workflows, patterns) with provenance metadata, verifies each via blind reconstruction, and shows that augmenting LLM‑based agents with these skills yields an average 11.7% performance lift across 72 proto…

    Hugging Face Daily Papersarxiv.org1 minpaper

  4. RiskChainBench: A Benchmark for Obfuscated Platform Message Restoration and Evidence-Grounded Web Investigation

    RiskChainBench is a new benchmark that pairs synthetic obfuscated message restoration inputs with human‑labeled local web environments, requiring models to both decode malicious instructions and investigate the linked site. Across ten models, restoration accuracy varies widely and web‑agent failures dominate the error budget.

    Hugging Face Daily Papersarxiv.org1 minpaper

  5. Dynamically Scaled Activation Steering

    Dynamically Scaled Activation Steering (DSAS) is a method‑agnostic framework that learns per‑token, per‑layer scaling factors to turn existing activation‑steering interventions on only when a model is likely to produce undesired output (e.g., toxic text). The scaling can be optimized jointly with any steering function, improves the toxicity‑utility trade‑off on language models, transfers to text‑…

    Apple Machine Learning Researchapple.com1 minpaper

  6. StepAudio 3 Realtime Technical Report

    StepAudio 3 Realtime is an audio‑language foundation model that runs a continuous listen‑converse‑think‑act loop. It introduces Deep Perception for rich acoustic cue extraction, Seamless Duplex for handling pauses/back‑channels, and a Think‑While‑Speaking mechanism that lets the model reason in parallel with speech output. On benchmarks it scores 73.0 macro avg on StepAudioChat, 90.6 on MMSU, 98.…

    Hugging Face Daily Papersarxiv.org2 minpaper