MEMRES: A Memory-Augmented Resolver with Confidence Cascade for Agentic Python Dependency Resolution
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917417941204992 |
|---|---|
| author | Minh, Dao Sy Duy Nguyen, Tran Chi Huynh, Trung Kiet Hoa, Pham Phu Quy, Nguyen Lam Phu Nguyen, Vu |
| author_facet | Minh, Dao Sy Duy Nguyen, Tran Chi Huynh, Trung Kiet Hoa, Pham Phu Quy, Nguyen Lam Phu Nguyen, Vu |
| contents | We present MEMRES, an agentic system for Python dependency resolution that introduces a multi-level confidence cascade where the LLM serves as the last resort. Our system combines: (1) a Self-Evolving Memory that accumulates reusable resolution patterns via tips and shortcuts; (2) an Error Pattern Knowledge Base with 200+ curated import-to-package mappings; (3) a Semantic Import Analyzer; and (4) a Python 2 heuristic detector resolving the largest failure category. On HG2.9K using Gemma-2 9B (10 GB VRAM). MEMRES resolves 2503 of 2890 (86.6%, 10-run average) snippets, combining intra-session memory with our confidence cascade for the remainder. This already exceeds PLLM's 54.7% overall success rate by a wide margin. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_16941 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MEMRES: A Memory-Augmented Resolver with Confidence Cascade for Agentic Python Dependency Resolution Minh, Dao Sy Duy Nguyen, Tran Chi Huynh, Trung Kiet Hoa, Pham Phu Quy, Nguyen Lam Phu Nguyen, Vu Software Engineering Artificial Intelligence We present MEMRES, an agentic system for Python dependency resolution that introduces a multi-level confidence cascade where the LLM serves as the last resort. Our system combines: (1) a Self-Evolving Memory that accumulates reusable resolution patterns via tips and shortcuts; (2) an Error Pattern Knowledge Base with 200+ curated import-to-package mappings; (3) a Semantic Import Analyzer; and (4) a Python 2 heuristic detector resolving the largest failure category. On HG2.9K using Gemma-2 9B (10 GB VRAM). MEMRES resolves 2503 of 2890 (86.6%, 10-run average) snippets, combining intra-session memory with our confidence cascade for the remainder. This already exceeds PLLM's 54.7% overall success rate by a wide margin. |
| title | MEMRES: A Memory-Augmented Resolver with Confidence Cascade for Agentic Python Dependency Resolution |
| topic | Software Engineering Artificial Intelligence |
| url | https://arxiv.org/abs/2604.16941 |