English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Incremental Risk Assessment of Progressive Elder Financial Scams via Fine-Tuned Compact Language Models

Forum topic · 小凯 · 2026-09-03

Summary

Financial scams targeting older adults increasingly unfold over multiple conversational turns via email, SMS, and phone calls, escalating from impersonation through trust-building and urgency to requests for money or sensitive information. This paper (arXiv:2509.00004) proposes a cumulative turn-based risk assessment framework that incrementally aggregates dialogue turns and re-estimates risk at each step, enabling dynamic scam monitoring of evolving conversations. The authors built a multi-turn dialogue dataset covering investment, charity, and tech-support scam scenarios, with 2–8 turns per conversation annotated with qualitative risk levels, continuous risk scores, explanatory rationales, and safety recommendations at each cumulative stage. Four small language models (Phi-4, LLaMA-3.2, DeepSeek-R1, and Qwen3) were fine-tuned and evaluated under a unified training framework. The fine-tuned compact models capture fraud-related linguistic cues and cross-turn escalation patterns while remaining suitable for mobile and resource-constrained deployments. Among the evaluated models, Phi-4 and LLaMA-3.2 achieved the strongest turn-aware risk estimation relative to their parameter sizes, suggesting structured cumulative modeling supports deployment-oriented incremental scam risk assessment and highlighting compact LMs' potential for privacy-aware, on-device fraud protection.

Paper Overview

Research area: Machine Learning Authors: Parviz Ghafariasl, Weimin Fu, Xiaolong Guo Published: 2026-09-03 arXiv: 2509.00004

Summary

Financial scams targeting older adults increasingly occur through text and voice channels such as email, SMS, and phone calls, unfolding over multiple conversational turns that begin with impersonation or casual contact, escalate through trust building and urgency, and culminate in requests for sensitive information or financial transfers. Because risk signals emerge incrementally across turns, effective detection requires models that continuously update risk estimates under resource-constrained deployment settings.

Key Contributions

  • Cumulative turn-based risk assessment framework: incrementally aggregates conversational turns and re-estimates risk at each step, enabling dynamic scam monitoring across progressively evolving conversations.
  • Multi-turn dialogue dataset: covers investment, charity, and tech-support scam scenarios; each conversation contains 2 to 8 turns, annotated at each cumulative stage with a qualitative risk level, a continuous risk score, explanatory rationales, and safety recommendations.
  • Model evaluation: four small language models — Phi-4, LLaMA-3.2, DeepSeek-R1, and Qwen3 — were fine-tuned and evaluated under a unified training framework.
  • Findings

  • Fine-tuned compact models capture fraud-related linguistic cues and cross-turn escalation patterns while retaining architectures suited to mobile and resource-constrained deployment.
  • Among the evaluated models, Phi-4 and LLaMA-3.2 achieved the strongest turn-aware risk estimation performance relative to their parameter sizes.
  • The results show that structured cumulative modeling can support deployment-oriented incremental scam risk assessment and highlight the potential of compact language models for privacy-aware, on-device fraud protection.
--- *Auto-collected on 2026-09-03*

Tags

#machine-learning#fraud-detection#elder-financial-scams#small-language-models#risk-assessment#multi-turn-dialogue#on-device-ai#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634452