llm-kb
← Back to research
Paper

Semantic Pareto-DQN: Multi-Objective RL for Financial Anomaly Detection

A new research paper proposes Semantic Pareto-DQN, a multi-objective reinforcement learning framework that uses LLMs to encode transaction features into natural-language narratives for financial anomaly detection. The method aims to overcome class imbalance and 'fraud collapse' without distortive data resampling.

0 engagement·1 source·Fri, Jul 10, 2026, 05:39 PM
The paper, titled 'Semantic Pareto-DQN: A Multi-Objective Reinforcement Learning Framework for Financial Anomaly Detection,' was published on arXiv on 2026-07-10. It addresses the problem of extreme class imbalance in financial anomaly detection, where traditional single-objective algorithms default to the majority class (fraud collapse) and fail to balance anomaly interdiction with customer friction. The proposed framework synthesizes heterogeneous transaction features into cohesive natural-language narratives encoded by large language models, producing a robust, scale-invariant state representation for the reinforcement learning agent. No benchmark results or model sizes are provided in the excerpt.

Entities

arXiv(tool)Semantic Pareto-DQN(concept)

Related

PaperFri, Jul 10, 2026, 10:19 AM

Paper proposes blockchain-linked LLM-based decision management for telecom fraud control

A new arXiv paper reframes telecom fraud control as blockchain-linked auditable decision management for synthetic telecom/IoT fraud-control requests. The main finding is that a QLoRA-tuned LLM branch becomes much more usable than zero-shot prompting but mainly approaches, rather than outperforms, a lower-cost centralized ensemble.

0 engagement·1 source·arxiv
arXiv
PaperThu, Jul 9, 2026, 06:00 PM

LLM/VLA models enable prompt-driven exploration in RL

A new research paper proposes using large language models (LLMs) and vision-language-action (VLA) models to drive exploration in reinforcement learning by modifying natural language prompts, which induce global behavioral changes beyond standard action noise. This approach could help policies escape weak local optima more effectively.

0 engagement·1 source·arxiv
arXiv
PaperFri, Jul 10, 2026, 12:42 AM

Paper proposes correlation-aware bandits with surrogate rewards for LLM routing

A new arXiv paper introduces contextual bandit algorithms that leverage surrogate reward signals from machine learning models to improve LLM routing decisions. The approach accounts for inter-arm correlations and noisy auxiliary rewards, addressing limitations of classical bandits that assume conditional independence.

0 engagement·1 source·arxiv
arXiv
PaperThu, Jul 9, 2026, 06:00 PM

Mixture of Probes: New method for multimodal LLMs to learn from privileged modalities

Researchers propose Mixture of Probes (MoP), a method enabling multimodal LLMs to leverage auxiliary modalities available only during training. This addresses real-world scenarios where inference-time modalities are limited, improving model performance without requiring all modalities at test time.

0 engagement·1 source·arxiv
arXiv
PaperThu, Jul 9, 2026, 09:53 PM

NL-PAC paper introduces framework for handling specification ambiguity in LLM supervision

A new arXiv paper proposes NL-PAC, a framework that addresses specification ambiguity when LLMs provide labels or evaluations from natural language instructions. The framework uses a fixed model's thresholded decoding law to define admissible labels and candidate targets, showing that additional labels reduce sampling error but cannot resolve identification problems from ambiguous specifications.

0 engagement·1 source·arxiv
arXiv