Negative Sampling Techniques in Information Retrieval: A Survey

2026-03-20T08:52:19Z01b51da48e9d4e50b4908458a64e46a0c92b75fee223232975a086530bac219f
API-integrationLLMMLOpsRAGauditabilitydata-exfiltrationdataset-contaminationevaluation-benchmarkshealthcare-safetymodel-biasmodel-hallucinationpatient-safetyprivacysynthetic-datatool-invocation

What happened

This collection of recent IR/LLM papers highlights several security- and safety-relevant developments: increased use of retrieval-augmented generation (RAG) and dynamic API/tool invocation (DynaRAG) which expand the attack surface (data exfiltration, malicious tool calls, supply-chain and third-party API risks); deterministic evidence-gating and sufficiency classifiers that improve auditability and reduce hallucinations; LLM-driven synthetic negative sampling and dataset synthesis that raise risks of data contamination and privacy leakage; healthcare-specific models (HypeMed) that, if misappli

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
arxiv_cs_ir
Record identifier
01b51da48e9d4e50b4908458a64e46a0c92b75fee223232975a086530bac219f
Enrichment time
2026-03-20T08:52:19Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.

Record · Negative Sampling Techniques in Information Retrieval: A Survey · Baitaphish