21 June 2021
Using domain randomization and meta-learning, computer vision models forget less when exposed to training samples from new domains. Remembering is a crucial element in the deployment of self-driving cars and robots which interact in dynamic environments.
2021
19 May 2021
NAVER LABS releases world's biggest visual localization dataset of indoor spaces with over 130K images. Dataset built with NAVER LABS mapping robots M1X & COMET and available in unified data format kapture.
2021
14 May 2021
Using a novel algorithm we explore how effectively a single policy, learned by reinforcement learning, can modulate robot behaviour from risk-averse to risk-neutral, so that robots can safely navigate everyday environments like homes and shops.
2021
30 April 2021
A novel framework for controlled NLG called 'Generation with Distributional Control', achieves great generality on the types of constraints that can be imposed and has a large potential to remedy the problem of bias in language models.
2021
20 April 2021
Papers and activities of NAVER LABS Europe at EACL 2021. NLG, scaling to larger context, cost of supervised data, controlling the output plus ASR at the AfricaNLP workshop.
2021
17 April 2021
Our Global BERT-based Transformer architecture fuses global and local information at every layer, resulting in a reading comprehension model that achieves a deeper understanding for long documents and enables flexibility for downstream tasks.
2021
31 March 2021
Visual localization is used by robots and self-driving cars to estimate their position and in AR apps to interact with the physical world both indoors and out in the open. This article gives an overview of current state-of-the-art methods and their advantages and drawbacks.
2021
10 March 2021
Podcast and transcript of Matthias Gallé, head of the NAVER research LAB in Europe who tells us what kind of research is going on in the labs in France and what it’s like to work there.
2021
27 February 2021
Agents are becoming more artificially intelligent but a number of problems must be solved before we can unleash them into our physical world (i.e. robots) and trust them in our digital one.
2021
18 December 2020
By dissecting the matching process of the recent ColBERT model, we make a step towards unveiling the ranking properties of BERT-based ranking models and show that ColBERT (implicitly) learns a notion of term importance that correlates with IDF.
2020
15 December 2020
Running into the unknown with RunAhead: finding pleasant running tours and intuitively navigating through them. The use of combinatorial optimization and UX design, together with head-tracking technology, provides a non-intrusive way of helping runners to discover new itineraries.
2020
9 December 2020
A novel framework which uses individual sample losses as error measures to determine the relative difficulty of samples in a dataset. Can be plugged on top of existing neural network models to implement curriculum learning for any task, even with noisy datasets.
2020