Showing 1–2 of 2 results for author: Hurtado, J F

Harnessing Retrieval-Augmented Generation (RAG) for Uncovering Knowledge Gaps

Authors: Joan Figuerola Hurtado

Abstract: The paper presents a methodology for uncovering knowledge gaps on the internet using the Retrieval Augmented Generation (RAG) model. By simulating user search behaviour, the RAG system identifies and addresses gaps in information retrieval systems. The study demonstrates the effectiveness of the RAG system in generating relevant suggestions with a consistent accuracy of 93%. The methodology can be… ▽ More The paper presents a methodology for uncovering knowledge gaps on the internet using the Retrieval Augmented Generation (RAG) model. By simulating user search behaviour, the RAG system identifies and addresses gaps in information retrieval systems. The study demonstrates the effectiveness of the RAG system in generating relevant suggestions with a consistent accuracy of 93%. The methodology can be applied in various fields such as scientific discovery, educational enhancement, research development, market analysis, search engine optimisation, and content development. The results highlight the value of identifying and understanding knowledge gaps to guide future endeavours. △ Less

Submitted 12 December, 2023; originally announced December 2023.

Attention Span For Personalisation

Authors: Joan Figuerola Hurtado

Abstract: A click on an item is arguably the most widely used feature in recommender systems. However, a click is one out of 174 events a browser can trigger. This paper presents a framework to effectively collect and store data from event streams. A set of mining methods is provided to extract user engagement features such as: attention span, scrolling depth and visible impressions. In this work, we presen… ▽ More A click on an item is arguably the most widely used feature in recommender systems. However, a click is one out of 174 events a browser can trigger. This paper presents a framework to effectively collect and store data from event streams. A set of mining methods is provided to extract user engagement features such as: attention span, scrolling depth and visible impressions. In this work, we present an experiment where recommendations based on attention span drove 340% higher click-through-rate than clicks. △ Less

Submitted 30 July, 2016; originally announced August 2016.