Web Science and Digital Libraries Research Group

As Natural Language Processing (NLP) evolves, Large Language Models (LLMs) are capable of interpreting text and extracting meaningful information. However, during my recent experimentation with Named Entity Recognition (NER) tasks, I found that different LLMs often produce varying outputs, much like humans offering diverse answers based on their unique perspectives. This leads to an interesting challenge: how can we unify the inconsistent entity list generated by multiple LLMs into a consistent and reliable output?  If you want to replicate these experiments, you can access the GitHub repository . Figure 1. Differing Entity Extraction Results from Two LLMs for the Same Text Input. The Challenge LLMs Capture Entities Differently For this experiment, I used Ollama , an open-source project that serves as a powerful and user-friendly platform for running LLMs on a local machine. Specifically, I employed Llama 3.1 8B (Meta), Phi 3.5 8B (Microsoft), and Gemma 2 9B (Google). During ...

  Project Aria glasses , introduced by Meta (formerly Facebook), represent a significant leap in wearable devices. Announced as part of Meta's larger vision for the metaverse , these glasses are designed to capture rich contextual egocentric multi-modal data, providing a platform for developing augmented reality (AR) experiences grounded in real-world interactions. Aria glasses are built to collect data about the environment and user behavior while opening new frontiers for eye tracking, spatial navigation, and human-computer interaction research. NirdsLab recently became a research partner for the program researching contextualized advanced eye tracking. In this blog post, we will discuss the capabilities of Aria glasses and how we use them in eye-tracking research. Figure 1: Components of the glasses ( https://facebookresearch.github.io ) At the current stage of development at Meta, the Aria glasses take the form of a sensor array for future smart glasses, containing visual ...