Intern, World Foundational Models Task-agnostic Latent Representations

Location: Rennes, France

About InterDigital

InterDigital is a global research and development company focused primarily on wireless, video, artificial intelligence (“AI”), and related technologies. We design and develop foundational technologies that enable connected, immersive experiences in a broad range of communications and entertainment products and services. We license our innovations worldwide to companies providing such products and services, including makers of wireless communications devices, consumer electronics, IoT devices, cars and other motor vehicles, and providers of cloud-based services such as video streaming. As a leader in wireless technology, our engineers have designed and developed a wide range of innovations that are used in wireless products and networks, from the earliest digital cellular systems to 5G and today’s most advanced Wi-Fi technologies. We are also a leader in video processing and video encoding/decoding technology, with a significant AI research effort that intersects with both wireless and video technologies. Founded in 1972, InterDigital is listed on Nasdaq.

InterDigital is a registered trademark of InterDigital, Inc.

For more information, visit: www.interdigital.com.

Summary

World Foundational Models (WFMs) are self-supervised AI models trained on vast datasets of sensors sequences (visual, sound, Lidar…) to learn the underlying physics, temporal dynamics, and spatial structures of physical environments.

Rather than predicting full-resolution, high-dimensional pixels for future states—which is computationally expensive and spends capacity on irrelevant noise like background texture changes—modern WFMs learn task-agnostic latent representations.

The goal of this internship is to explore, benchmark, and evaluate how task-agnostic latent spaces learned by existing World Foundational Models (like NVIDIA Cosmos, Meta’s V-JEPA or LeWorldMode) can infer information and interpret that information to take decisions. One expected outcome is to demonstrate that the WFMs are generic enough to infer different information for different applications.

Target use cases may include autonomous navigation, elderly people surveillance, and anomaly detection. The work may lead to prototype software, internal demonstrators, and potentially publication and/or patent contributions.

Responsibilities

The intern will join the haptics research team within the New Media Codec group at InterDigital (Rennes). The work will consist in:

Studying the state of the art on neural representations of world models

Evaluate and compare different models

Designing and implementing applications using those models and identify the type of inferences that can be expected

Exploring agents able to use those inferences for some applications (autonomous robots, alert detection, elderly care…)

Evaluating the proposed methods and developing demonstrators with potential contributions to publications and patents

Qualifications

Master’s student (or 4th year engineering school) in Computer Science, Machine Learning, Signal Processing, Robotics, or related field

Strong Python programming skills and experience with PyTorch or TensorFlow

Good knowledge of neural networks, representation learning, and signal processing

Nice to have

Experience with autoencoders, GANs, INRs, or generative models for temporal data

Knowledge of video, audio, time-series processing, or interactive systems

Experience with C++, real-time systems, or game engines

Keywords: Neural Networks, Generative Models, Autoencoders, GANs, INR, Representation Learning, Procedural Generation, Autonomous systems

Expected Outcomes: Prototype software, demos, potentially publications and patents

Mentors: Gurvan Lécuyer, Quentin Galvane

InterDigital is an equal employment opportunity employer. InterDigital will not engage in or tolerate unlawful discrimination with regard to any employment decision, policy or practice based on a person’s sex, gender, pregnancy (including childbirth, breastfeeding and related medical conditions), age, race, color, religion, creed, national origin, ancestry, citizenship, military status, veteran status, mental or physical disability, medical condition, genetic information, sexual orientation, gender identity or expression, or any other factor protected by applicable federal, state or local  law. This policy applies to all terms and conditions of employment, including, but not limited to, recruiting, hiring, compensation, benefits, training, assignments, evaluations, coaching, promotion, discipline, discharge and layoff.

À propos d'InterDigital

InterDigital est une société internationale de recherche et développement spécialisée principalement dans les technologies sans fil, la vidéo, l'intelligence artificielle (« IA ») et les technologies associées. Nous concevons et développons des technologies fondamentales qui rendent possibles des expériences connectées et immersives dans un large éventail de produits et services de communication et de divertissement. Nous concédons des licences sur nos innovations dans le monde entier à des entreprises qui proposent ces produits et services, notamment des fabricants d'appareils de communication sans fil, d'électronique grand public, d'objets connectés (IoT), d'automobiles et d'autres véhicules à moteur, ainsi qu'à des fournisseurs de services cloud tels que la diffusion vidéo en streaming. Leader des technologies sans fil, InterDigital a vu ses ingénieurs concevoir et développer un large éventail d'innovations utilisées dans les produits et réseaux sans fil, des premiers systèmes cellulaires numériques jusqu'à la 5G et aux technologies Wi-Fi les plus avancées d'aujourd'hui. Nous sommes également un leader du traitement vidéo et des technologies de codage et de décodage vidéo, avec un important effort de recherche en IA à la croisée des technologies sans fil et vidéo. Fondée en 1972, InterDigital est cotée au Nasdaq.

Int