iSemantics's repositories
Finance-SC-Relations
Finance-SC-Relations is a project that focuses on extracting supply-chain relations from financial text using Transformer-based models and Large Language Models (LLMs). The project aims to enhance the accuracy and reliability of relation extraction in the supply-chain domain by leveraging the power of Transformer based models.
Entity-Linking
Entity linking is a project that involves using rule-based normalization methods developed in Inferess to create pairs of positive and hard negative datasets. These datasets are then used to train a contrastive encoder capable of mapping company names into 384 vectors. The matching process is performed using the Faiss Algorithm.
Language:Jupyter NotebookApache-2.0000