Legal Case Retrieval (LCR), which retrieves relevant cases from a query case, is a fundamental task for legal professionals in research and decision-making.However, existing studies on LCR face two major limitations.First, they are evaluated on relatively small-scale retrieval corpora (e.g., 100-55K cases) and use a narrow range of criminal query types, which cannot sufficiently reflect the complexity of real-world legal retrieval scenarios.Second, their reliance on embeddingbased or lexical matching methods often results in limited representations and legally irrelevant matches.To address these issues, we present: (1) LEGAR BENCH, the first largescale Korean LCR benchmark, covering 411 diverse crime types in queries over 1.2M candidate cases; and (2) LegalSearchLM, a retrieval model that performs legal element reasoning over the query case and directly generates content containing those elements, grounded in the target cases through constrained decoding.Experimental results show that LegalSearchLM outperforms baselines by 6 -20% on LEGAR BENCH, achieving state-of-the-art performance.It also demonstrates strong generalization to out-of-domain cases, outperforming naive generative models trained on in-domain data by 15%.

Seungone Kim, Juyoung Suk, Ji Yong Cho, Shayne Longpre, Chaeeun Kim, Dongkeun Yoon, Guijin Son, Yejin Cho, Sheikh Shafayat, Jinheon Baek, Sue Hyun Park, Hyeonbin Hwang, Jinkyung Jo, Hyowon Cho, Haebin Shin, Seongyun Lee, Hanseok Oh, Noah Lee, Namgyu Ho, Se June Joo, Miyoung Ko, Yoonjoo Lee, Hyungjoo Chae, Jamin Shin, Joel Jang, Seonghyeon Ye, Bill Yuchen Lin, Sean Welleck, Graham Neubig, Moontae Lee, Kyungjae Lee, Minjoon Seo. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Edge applications generate a large influx of sensor data on massive scales, and these massive data streams must be processed shortly to derive actionable intelligence. However, traditional data processing systems are not well-suited for these edge applications as they often do not scale well with a large number of concurrent stream queries, do not support low-latency processing under limited edge computing resources, and do not adapt to the level of heterogeneity and dynamicity commonly present in edge computing environments. As such, we present AgileDart, an agile and scalable edge stream processing engine that enables fast stream processing of many concurrently running low-latency edge applications' queries at scale in dynamic, heterogeneous edge environments. The novelty of our work lies in a dynamic dataflow abstraction that leverages distributed hash table-based peer-to-peer overlay networks to autonomously place, chain, and scale stream operators to reduce query latencies, adapt to workload variations, and recover from failures and a bandit-based path planning model that re-plans the data shuffling paths to adapt to unreliable and heterogeneous edge networks. We show that AgileDart outperforms Storm and EdgeWise on query latency and significantly improves scalability and adaptability when processing many real-world edge stream applications' queries.

Benchmarking the performance of information retrieval (IR) is mostly conducted with a fixed set of documents (static corpora).However, in realistic scenarios, this is rarely the case and the documents to be retrieved are constantly updated and added.In this paper, we focus on Generative Retrievals (GR), which apply autoregressive language models to IR problems, and explore their adaptability and robustness in dynamic scenarios.We also conduct an extensive evaluation of computational and memory efficiency, crucial factors for real-world deployment of IR systems handling vast and ever-changing document collections.Our results on the StreamingQA benchmark demonstrate that GR is more adaptable to evolving knowledge (4 -11%), robust in learning knowledge with temporal information, and efficient in terms of inference FLOPs (ˆ2), indexing time (ˆ6), and storage footprint (ˆ4) compared to Dual Encoders (DE), which are commonly used in retrieval systems.Our paper highlights the potential of GR for future use in practical IR systems within dynamic environments.

How can we efficiently find connected components with Union-Find in a distributed system? Union-Find is the most efficient sequential algorithm for finding connected components with low memory usage and high speed. Several studies have adapted Union-Find to distributed memory systems to process large graphs quickly; however, they all suffer from load balancing problems. We notice that the leading cause of the load balancing problems is the nature of Union-Find, which gathers more and more edges to a small number of vertices as it proceeds. In this paper, we propose BTS, a new fast and scalable distributed Union-Find algorithm for finding connected components in large graphs. BTS resolves the load balancing problems by proposing Balanced Union-Find, which allocates vertices to each processor and makes edges link to vertices in the same processor as much as possible. We further optimize BTS with edge refinement to minimize network traffic and memory usage. Experimental results show that BTS efficiently resolves the load balancing problems, processing 16–1024 times larger graphs with 3.1-261.9 times faster speeds than existing algorithms.

Hyunji Lee, Se June Joo, Chaeeun Kim, Joel Jang, Doyoung Kim, Kyoung-Woon On, Minjoon Seo. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.