Pith. sign in

REVIEW 1 cited by

Knowledge Distillation in Document Retrieval

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1911.11065 v1 pith:R44UDUSF submitted 2019-11-11 cs.IR cs.CLcs.LG

Knowledge Distillation in Document Retrieval

classification cs.IR cs.CLcs.LG
keywords documentclaimmodelscomplexindependentmodelretrievaldistillation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X LinkedIn Reddit HN
read the original abstract

Complex deep learning models now achieve state of the art performance for many document retrieval tasks. The best models process the query or claim jointly with the document. However for fast scalable search it is desirable to have document embeddings which are independent of the claim. In this paper we show that knowledge distillation can be used to encourage a model that generates claim independent document encodings to mimic the behavior of a more complex model which generates claim dependent encodings. We explore this approach in document retrieval for a fact extraction and verification task. We show that by using the soft labels from a complex cross attention teacher model, the performance of claim independent student LSTM or CNN models is improved across all the ranking metrics. The student models we use are 12x faster in runtime and 20x smaller in number of parameters than the teacher

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Enhancing deep learning models for time series classification via knowledge distillation

    cs.LG 2026-07 conditional novelty 4.0

    Knowledge distillation most benefits intermediate-complexity students for time series classification, cutting parameters sharply while matching teacher accuracy across FCN, Inception, and ConvTran on UCR.