---
title: "NIR: Approximate Neural Instance Retrieval for ALCIQ"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.02842v1"
permalink: "https://feed.astralor.com/items/itm_867b7f0da8e406c3d0744e8f"
published: "2026-11-02T09:44:18.000Z"
score: 53
channel: "tech"
tags: ["RAG","NIR","Approximate Neural Instance","Retrieval","ALCIQ Instance","ALCIQ","Instance"]
---

# NIR: Approximate Neural Instance Retrieval for ALCIQ

## 速览

Instance retrieval is the main bottleneck of concept-learning systems based on refinement operators. We present NIR, a family of neural instance retrievers for ALCIQ that output graded membership scores from knowledge g…

## 关键点

- Instance retrieval is the main bottleneck of concept-learning systems based on refinement operators.
- We present NIR, a family of neural instance retrievers for ALCIQ that output graded membership scores from knowledge graph embeddings.
- We compare tokenization-based encoders (GRU, LSTM, Transformer) with a Compositional encoder that mirrors DL operators via primitive sub-networks.

## 全文

Instance retrieval is the main bottleneck of concept-learning systems based on refinement operators. We present NIR, a family of neural instance retrievers for ALCIQ that output graded membership scores from knowledge graph embeddings. We compare tokenization-based encoders (GRU, LSTM, Transformer) with a Compositional encoder that mirrors DL operators via primitive sub-networks. On eight benchmarks, all variants approximate Pellet closely on complete knowledge bases, while the Transformer model is most robust under incompleteness. The source code of our approach is at https://github.com/dice-group/owlapy.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_867b7f0da8e406c3d0744e8f