2 papers
cs.IR2026
Understanding and Debugging Failures in N-Gram-Based Generative Retrieval
Richard Takacs, Adrian Bracher, Svitlana Vakulenko
Generative Retrieval (GR) is an emerging Information Retrieval (IR) paradigm that is motivated by increasingly capable language models. In GR, a model directly generates identifier…
cs.IR2026
Generative Retrieval Overcomes Limitations of Dense Retrieval but Struggles with Identifier Ambiguity
Adrian Bracher, Svitlana Vakulenko
While dense retrieval models, which embed queries and documents into a shared low-dimensional space, have gained widespread popularity, they were shown to exhibit important theoret…