Papers Club: Graph Reasoning meets Language Understanding

We are excited to announce our upcoming Papers Club session where we delve into the fascinating intersection of Graph Neural Networks (GNNs) and Language Models (LMs) in processing Text Attributed Graphs (TAGs).

Graphs, such as knowledge graphs, often contain text attributes in their nodes and edges, making them a complex challenge for traditional models. Language Models are great at handling text but struggle with the reasoning required for graph structures. On the other hand, Graph Neural Networks excel at reasoning over graphs but can’t process text as effectively.

In this session, Moritz Platz will explore methods for reasoning over TAGs, discussing the strengths and limitations of each approach. We’ll also cover strategies for interleaving text and graph inputs, enabling models to jointly reason across both modalities.

Key takeaways:
📌 Overview of the latest research in Graph Language Models
📌 Insights into how models can combine text and graph reasoning
📌 Discussion of multiple methods for working with Text Attributed Graphs

Join us for this deep dive into cutting-edge AI techniques that blend natural language understanding with graph reasoning!

By registering for the event, you consent to the processing of your data by Roosh LTD, its affiliates, partners, and co-hosts.