๐ Excited to share our new paper accepted at ICML 2026!
๐ Did you know retrievers can have their own blind spots-where certain documents are consistently hard to retrieve?
๐ Retrieval failures are not random; some documents are consistently hard to retrieve.
๐ We introduce RPS to quantify retrievability and show it can be predicted from embeddings before indexing.
๐ ๏ธ ARGUS detects these blind spots and mitigates them with targeted augmentation.
๐ We observe consistent gains in retrieval performance (+3.4 nDCG@5, +4.5 nDCG@10 on average).
๐ค Grateful to work with an amazing team: @AModarressi , @HinrichSchuetze , and @amarfurt, at @CisLmu and @hslu.
๐ป Code, supplementary materials, and reimplementations: https://t.co/knvjqY1uaO
๐ Paper: https://t.co/6Y68KIe3Dn
#ICML2026
#ICML
#Trustworthy
#RAG
#InformationRetrieval