arXiv subjects

Minjun Choi

Publications and source records attributed to Minjun Choi.

3 recordsLinked to original sources

arXiv2026

KG2Cypher: Data-Centric Pipeline for Building Enterprise Text-to-Cypher Systems

Enterprise Knowledge Graphs (KGs) are increasingly used for internal search, analytics, and question answering, but building natural-language interfaces for private enterprise graphs remains costly. We present KG2Cypher, a data-centric pipeline for building enterprise text-to-Cypher systems from existing KGs. KG2Cypher first constructs an executable Cypher query from observed graph facts and then uses LLMs to generate its associated natural-language question. The resulting Text-Cypher pairs are validated with an LLM judge and human validation, and are converted into candidate-aware SFT data. The trained generator is served with class-conditioned schema prompting, entity retrieval, and LoRA-based inference. We evaluate KG2Cypher in Korean enterprise settings, where short search-style queries and schema paraphrases make language grounding difficult. LoRA SFT improves execution-result F1 from 0.806 to 0.950 on broadcast-program queries and from 0.70 to 0.92 on company queries. In an 11-class setting, KG2Cypher achieves 95.2% exact match, 99.9% execution rate, and 0.964 execution-result F1.

cs.CL↗

arXiv2026

G-Long: Graph-Enhanced Memory Management for Efficient Long-Term Dialogue Agents

While Large Language Models (LLMs) have advanced open-domain dialogue systems, maintaining long-term consistency remains a challenge due to inherent limitations in long-context reasoning and the inefficiency of processing extensive raw text. Existing approaches typically rely on either unstructured memory storage, which is prone to information loss, or computationally expensive LLMs that incur high latency. To address these limitations, we propose G-Long, a graph-enhanced framework that utilizes a fine-tuned small Language Model (sLM) for structured triplet extraction and associative retrieval, significantly reducing operational costs. Furthermore, we introduce the novel attention-aware importance scoring mechanism that leverages the intrinsic cross-attention signals of a T5 summarizer to identify salient memories. Extensive experiments across diverse benchmarks demonstrate that G-Long achieves state-of-the-art performance in both response generation and memory retrieval, yielding performance gains of up to 9.8% in response quality on MSC and 40.8% in retrieval recall on LME, while significantly minimizing computational overhead.

cs.CL↗

arXiv2021

Near real-time streaming analysis of big fusion data

While experiments on fusion plasmas produce high-dimensional data time series with ever increasing magnitude and velocity, data analysis has been lagging behind this development. For example, many data analysis tasks are often performed in a manual, ad-hoc manner some time after an experiment. In this article we introduce the DELTA framework that facilitates near real-time streaming analysis of big and fast fusion data. By streaming measurement data from fusion experiments to a high-performance compute center, DELTA allows to perform demanding data analysis tasks in between plasma pulses. This article describe the modular and expandable software architecture of DELTA and presents performance benchmarks of its individual components as well as of entire workflows. Our focus is on the streaming analysis of ECEi data measured at KSTAR on NERSCs supercomputers and we routinely achieve data transfer rates of about 500 Megabyte per second. We show that a demanding turbulence analysis workload can be distributed among multiple GPUs and executes in under 5 minutes. We further discuss how DELTA uses modern database systems and container orchestration services to provide web-based real-time data visualization. For the case of ECEi data we demonstrate how data visualizations can be augmented with outputs from machine learning models. By providing session leaders and physics operators results of higher order data analysis using live visualization they may monitor the evolution of a long-pulse discharge in near real-time and may make more informed decision on how to configure the machine for the next shot.

physics.plasm-ph↗