Category: News

FraudShield: Knowledge Graph Empowered Defense for LLMs against Fraud Attackscs.AI updates on arXiv.org arXiv:2601.22485v1 Announce Type: cross
Abstract: Large language models (LLMs) have been widely integrated into critical automated workflows, including contract review and job application processes. However, LLMs are susceptible to manipulation by fraudulent information, which can lead to harmful outcomes. Although advanced defense methods have been developed to address this issue, they often exhibit limitations in effectiveness, interpretability, and generalizability, particularly when applied to LLM-based applications. To address these challenges, we introduce FraudShield, a novel framework designed to protect LLMs from fraudulent content by leveraging a comprehensive analysis of fraud tactics. Specifically, FraudShield constructs and refines a fraud tactic-keyword knowledge graph to capture high-confidence associations between suspicious text and fraud techniques. The structured knowledge graph augments the original input by highlighting keywords and providing supporting evidence, guiding the LLM toward more secure responses. Extensive experiments show that FraudShield consistently outperforms state-of-the-art defenses across four mainstream LLMs and five representative fraud types, while also offering interpretable clues for the model’s generations.

 arXiv:2601.22485v1 Announce Type: cross
Abstract: Large language models (LLMs) have been widely integrated into critical automated workflows, including contract review and job application processes. However, LLMs are susceptible to manipulation by fraudulent information, which can lead to harmful outcomes. Although advanced defense methods have been developed to address this issue, they often exhibit limitations in effectiveness, interpretability, and generalizability, particularly when applied to LLM-based applications. To address these challenges, we introduce FraudShield, a novel framework designed to protect LLMs from fraudulent content by leveraging a comprehensive analysis of fraud tactics. Specifically, FraudShield constructs and refines a fraud tactic-keyword knowledge graph to capture high-confidence associations between suspicious text and fraud techniques. The structured knowledge graph augments the original input by highlighting keywords and providing supporting evidence, guiding the LLM toward more secure responses. Extensive experiments show that FraudShield consistently outperforms state-of-the-art defenses across four mainstream LLMs and five representative fraud types, while also offering interpretable clues for the model’s generations. Read More  

NVIDIA AI Brings Nemotron-3-Nano-30B to NVFP4 with Quantization Aware Distillation (QAD) for Efficient Reasoning InferenceMarkTechPost NVIDIA has released Nemotron-Nano-3-30B-A3B-NVFP4, a production checkpoint that runs a 30B parameter reasoning model in 4 bit NVFP4 format while keeping accuracy close to its BF16 baseline. The model combines a hybrid Mamba2 Transformer Mixture of Experts architecture with a Quantization Aware Distillation (QAD) recipe designed specifically for NVFP4 deployment. Overall, it is an ultra-efficient
The post NVIDIA AI Brings Nemotron-3-Nano-30B to NVFP4 with Quantization Aware Distillation (QAD) for Efficient Reasoning Inference appeared first on MarkTechPost.

 NVIDIA has released Nemotron-Nano-3-30B-A3B-NVFP4, a production checkpoint that runs a 30B parameter reasoning model in 4 bit NVFP4 format while keeping accuracy close to its BF16 baseline. The model combines a hybrid Mamba2 Transformer Mixture of Experts architecture with a Quantization Aware Distillation (QAD) recipe designed specifically for NVFP4 deployment. Overall, it is an ultra-efficient
The post NVIDIA AI Brings Nemotron-3-Nano-30B to NVFP4 with Quantization Aware Distillation (QAD) for Efficient Reasoning Inference appeared first on MarkTechPost. Read More  

“Existential risk” – Why scientists are racing to define consciousnessArtificial Intelligence News — ScienceDaily Scientists warn that rapid advances in AI and neurotechnology are outpacing our understanding of consciousness, creating serious ethical risks. New research argues that developing scientific tests for awareness could transform medicine, animal welfare, law, and AI development. But identifying consciousness in machines, brain organoids, or patients could also force society to rethink responsibility, rights, and moral boundaries. The question of what it means to be conscious has never been more urgent—or more unsettling.

 Scientists warn that rapid advances in AI and neurotechnology are outpacing our understanding of consciousness, creating serious ethical risks. New research argues that developing scientific tests for awareness could transform medicine, animal welfare, law, and AI development. But identifying consciousness in machines, brain organoids, or patients could also force society to rethink responsibility, rights, and moral boundaries. The question of what it means to be conscious has never been more urgent—or more unsettling. Read More  

How to Run Claude Code for Free with Local and Cloud Models from OllamaTowards Data Science Ollama now offers Anthropic API compatibility
The post How to Run Claude Code for Free with Local and Cloud Models from Ollama appeared first on Towards Data Science.

 Ollama now offers Anthropic API compatibility
The post How to Run Claude Code for Free with Local and Cloud Models from Ollama appeared first on Towards Data Science. Read More  

How to Apply Agentic Coding to Solve ProblemsTowards Data Science Learn how to efficiently solve problems with coding agents
The post How to Apply Agentic Coding to Solve Problems appeared first on Towards Data Science.

 Learn how to efficiently solve problems with coding agents
The post How to Apply Agentic Coding to Solve Problems appeared first on Towards Data Science. Read More