The field of natural language processing is witnessing significant advancements with the integration of large language models (LLMs) and knowledge graphs (KGs). Recent developments indicate a shift towards enhancing the capabilities of LLMs through external tools and knowledge graphs to improve accuracy and performance. The focus is on creating complex computing ecosystems around LLMs to support various tasks and activities. Noteworthy papers in this regard include the introduction of the Athena framework, which achieves state-of-the-art results in mathematical and scientific reasoning, and the KG-Attention framework, which enables dynamic knowledge fusion without parameter updates. Another significant development is the proposal of the DuetGraph mechanism, which tackles over-smoothing in KG reasoning and achieves state-of-the-art performance. The benefits of query-based KGQA systems for complex and temporal questions are also being explored, with promising results. Additionally, the development of specialized benchmarks such as Ref-Long and BOOKCOREF is helping to assess the capabilities of LLMs in long-context understanding and coreference resolution. Overall, the field is moving towards more sophisticated and effective integration of LLMs and KGs to enhance performance and accuracy in various NLP tasks.
Advancements in Large Language Models and Knowledge Graphs
Sources
Integrating External Tools with Large Language Models to Improve Accuracy
GRASP: Generic Reasoning And SPARQL Generation across Knowledge Graphs
KG-Attention: Knowledge Graph-Guided Attention at Test-Time via Bidirectional Information Aggregation
S2SRec2: Set-to-Set Recommendation for Basket Completion with Recipe
OPENXRD: A Comprehensive Benchmark and Enhancement Framework for LLM/MLLM XRD Question Answering
Ref-Long: Benchmarking the Long-context Referencing Capability of Long-context Language Models
Prompting for Performance: Exploring LLMs for Configuring Software
Automating SPARQL Query Translations between DBpedia and Wikidata
Introducing the Swiss Food Knowledge Graph: AI for Context-Aware Nutrition Recommendation
Enhancing the Capabilities of Large Language Models for API calls through Knowledge Graphs
Semantic Context for Tool Orchestration
WhisperKit: On-device Real-time ASR with Billion-Scale Transformers
DuetGraph: Coarse-to-Fine Knowledge Graph Reasoning with Dual-Pathway Global-Local Fusion
The benefits of query-based KGQA systems for complex and temporal questions in LLM era
BOOKCOREF: Coreference Resolution at Book Scale
Improving Contextual ASR via Multi-grained Fusion with Large Language Models
A Conceptual Framework for Requirements Engineering of Pretrained-Model-Enabled Systems
A Survey of Context Engineering for Large Language Models