Structure & Knowledge Injection into Machine Learning
Embedded Files
Recent News
Recent News
11/15/25 "A survey on large language models in biology and chemistry" accepted at EMM 2025!
11/08/25 "Latent Self-Consistency for Reliable Majority" accepted at AAAI 2026!
10/06/25 "Stop-RAG: Value-Based Retrieval Control for Iterative RAG" accepted at NeurIPS 2025 workshop!
08/20/25 "GraphCheck: Multi-Path Fact-Checking with Entity-Relationship Graphs" accepted at EMNLP 2025!
08/20/25 "SEAL-Pose: Enhancing 3D Human Pose Estimation through Trainable Loss Function" accepted at ICCV 2025 workshop!
06/30/25 4 interns have joined SKI-ML.
06/12/25 5 graduate students have joined SKI-ML.
05/21/25 "Set consistency verificaiton (Oral)" & "BridG MT: Enhancing LLM MT through sentence bridging" papers accepted at ACL 2025!
03/13/25 1 intern has joined SKI-ML.
01/23/25 "CoMRes: Semi-Supervised Time Series Forecasting Utilizing Consensus Promotion of Multi-Resolution" accepted at ICLR 2025!
01/13/25 1 intern has joined SK-ML.
12/01/24 8 graduate students have joined SKI-ML.
11/29/24 1 paper accepted at COLING 2025 (Oral)!
Research Topics
Research Topics
NLP Developing AI models that use large language models to enhance their ability to perform complex reasoning tasks.
NLP Developing AI models that use large language models to enhance their ability to perform complex reasoning tasks.
Knowledge Injection Focusing on the integration of human knowledge into machine learning models to enhance learning with constraints.
Knowledge Injection Focusing on the integration of human knowledge into machine learning models to enhance learning with constraints.
Structured Prediction Leveraging tools such as Energy-Based Models (EBM) to predict complex structured data.
Structured Prediction Leveraging tools such as Energy-Based Models (EBM) to predict complex structured data.
SKI-ML (Structure and Knowledge Injection into Machine Learning) aims to enhance machine learning capabilities by incorporating structured prediction and human intelligence into AI models. Our lab specializes in the strategic use of advanced machine learning techniques such as Energy-Based Models (EBM) to process and predict structured data efficiently. We engage in various NLP tasks using large language models (LLM), with a particular emphasis on improving their reasoning abilities. This integrative approach not only boosts the adaptability and precision of AI systems but also deepens the synergy between human knowledge and artificial intelligence, enabling our technologies to excel in diverse real-world applications. Our work aims to bridge theoretical insights and practical implementations, propelling a deeper connection and understanding between the realms of human cognition and machine learning.
SKI-ML (Structure and Knowledge Injection into Machine Learning) aims to enhance machine learning capabilities by incorporating structured prediction and human intelligence into AI models. Our lab specializes in the strategic use of advanced machine learning techniques such as Energy-Based Models (EBM) to process and predict structured data efficiently. We engage in various NLP tasks using large language models (LLM), with a particular emphasis on improving their reasoning abilities. This integrative approach not only boosts the adaptability and precision of AI systems but also deepens the synergy between human knowledge and artificial intelligence, enabling our technologies to excel in diverse real-world applications. Our work aims to bridge theoretical insights and practical implementations, propelling a deeper connection and understanding between the realms of human cognition and machine learning.
Page updated
Google Sites
Report abuse