Pierre Muletier

Pierre is a French/Italian Engineer born and raised in Switzerland.

Passionate from a young age about art and music just as much as physics and maths, he pursued a double Master’s degree at the Royal College of Art and Imperial College London, with the goal to bridge creative work with the technical and scientific rigour he had learned during his BEng. Since then, his projects have included a space-time curvature inspired armchair, CreAItive - an ML-powered creative collaborator, Grove - a sustainable social media platform and lastly, Paperceive – a research platform aiming to democratise academic knowledge.

In his free time, Pierre enjoys drawing, playing the drums, and taking online ML and data science courses.

Paperceive is an online tool that makes academic research fast and easy. Find papers intuitively through a visual search engine and understand their contents thanks to AI-powered text analysis.

The digitalisation of academic publications has made scientific knowledge more accessible than ever, but for most, academic research remains a slow, laborious, and frustrating process.

Finding the “perfect paper” can be very challenging, as current search engines rely on the usage of very precise search terms to successfully filter the 400+ million existing publications. Their list-based user-interface interface completely deprives the user of the broader context they are researching. Even when relevant literature is found, understanding its contents can still be problematic as it is often written using a highly complex language and style.

Paperceive overcomes these challenges by accompanying users in in their research journey, from the discovery of papers (by presenting search results as a visual network) to their understanding (leveraging state of the art machine-learning models to generate highly accurate and comprehensive, plain English translations).

Through expert validation and extensive user testing, the platform has proven its inherent value not only for university students but hopefully one day for any individual who seeks to explore, learn, and innovate.

The Problem

Paperceive presents a visual map of the area you are researching based on your search terms and neighbouring keywords it has identified within academic literature.

The contextual information, where similar papers are automatically displayed closer to each other, enables the user to visually pinpoint the area most relevant to their research, taking them intuitively to the most suitable paper. This overcomes the reliance on obscure keywords while increasing potential serendipity during the research process.

Hyper Summary

Hyper Summary runs on a custom trained language model, which uses machine-learning to understand the semantics of a selected text, and generate a clear and concise “translation” in plain English. The feature is powered by OpenAI's "GPT3", the single most powerful natural-language processing model available today.

The accuracy of the AI-generated summaries was validated by published researchers in the fields of medicine, material science, and machine learning.

To find out more about other features and how the platform works, or simply join me on my mission to democratise academic knowledge, don't hesitate to contact me!

LinkedIn: https://www.linkedin.com/in/pierremuletier/

e-mail: pierre@muletier.com

Pierre Muletier

Social

About

Royal College of Art