My submission for the research track at the 2022 International Semantic Web Conference, titled “EaT-PIM: Substituting Entities in Procedural Instructions using Flow Graphs and Embeddings”, was accepted. This paper is based on the work that I conducted during my internship at Bosch last year, and consists of methods to parse natural language instructions into flow graphs as well as a novel embedding method which we can use to then identify plausible ingredient substitions.

The repo of the code used in this work, as well as a copy of the camera ready version of our paper, is available here.

At ISWC 2022, I also had a number of additional papers that I was involved with accepted.

First, at the Wikidata Workshop I presented our work on Rule-Based Link Prediction over Event-Related Causal Knowledge in Wikidata. In this work, we curated a dataset of events with causal relations in Wikidata, and investigated the performance of some rule-based link prediction methods over this data.

At the DL4KG workshop, a fellow IBM intern Aamod shared our investigations on Knowledge Graph Embeddings for Causal Relation Prediction, in which we investigate the performance of various SotA link prediction models over another dataset of Wikidata events and their causal relations.

Lastly, I also contributed to some work in an in-use track submission surrounding IBM’s Knowledge-Based News Event Analysis Toolkit.