Finding Needles in Emb(a)dding Haystacks: Legal Document Retrieval via Bagging and SVR Ensembles

Kevin Bönisch, Alexander Mehler
1/9/2025
cs.IRcs.AI

No official code found. Best implementation: TheItCrOw/lirai24 (3 stars, last commit 2025-01-10).

Abstract

We introduce a retrieval approach leveraging Support Vector Regression (SVR) ensembles, bootstrap aggregation (bagging), and embedding spaces on the German Dataset for Legal Information Retrieval (GerDaLIR). By conceptualizing the retrieval task in terms of multiple binary needle-in-a-haystack subtasks, we show improved recall over the baselines (0.849 > 0.803 | 0.829) using our voting ensemble, suggesting promising initial results, without training or fine-tuning any deep learning models. Our approach holds potential for further enhancement, particularly through refining the encoding models and optimizing hyperparameters.

View on arXivView PDF

Code Implementations(1)

TheItCrOw/lirai24Community100%

Finding Needles in Emb(a)dding Haystacks: Legal Document Retrieval via Bagging and SVR Ensembles

30Jupyter NotebookJul 3, 20241 years ago
aidata-scienceembeddingsinformation-retrievallegal+1 more

Similar Papers

Test-time regression: a unifying framework for designing sequence models with associative memory

79% match2025

Guiding Retrieval using LLM-based Listwise Rankers

78% match1 repo32025

SPENCER: Self-Adaptive Model Distillation for Efficient Code Retrieval

78% match2025

Unifying Generative and Dense Retrieval for Sequential Recommendation

78% match1 repo432024

Catching Image Retrieval Generalization

77% match2023

From Data Deluge to Data Curation: A Filtering-WoRA Paradigm for Efficient Text-based Person Search

77% match1 repo92024

Cite this paper

@article{bnisch2025finding,
  title  = {Finding Needles in Emb(a)dding Haystacks: Legal Document Retrieval via Bagging and SVR Ensembles},
  author = {Kevin Bönisch and Alexander Mehler},
  year   = {2025},
  eprint = {2501.05018},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2501.05018v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2501.05018.svg)](https://codeofpaper.com/paper/2501.05018)

Discussion