Publications (23)
Atlas: Few-shot Learning with Retrieval Augmented Language Models
Gautier Izacard, Patrick Lewis, Maria Lomeli +7
Large language models have shown impressive few-shot results on a wide range of tasks. However, when knowledge is key for such results, as is the case for tasks such as question an…
Improving Wikipedia Verifiability with AI
Fabio Petroni, Samuel Broscheit, Aleksandra Piktus +10
Verifiability is a core content policy of Wikipedia: claims that are likely to be challenged need to be backed by citations. There are millions of articles available online and tho…
Stochastic activations
Maria Lomeli, Matthijs Douze, Gergely Szilvasy +7
We introduce stochastic activations. This novel strategy randomly selects between several non-linear functions in the feed-forward layer of a large language model. In particular, w…
Universal Marginalizer for Amortised Inference and Embedding of Generative Models
Robert Walecki, Albert Buchard, Kostis Gourgoulias +6
Probabilistic graphical models are powerful tools which allow us to formalise our knowledge about the world and reason about its inherent uncertainty. There exist a considerable nu…
General Latent Feature Models for Heterogeneous Datasets
Isabel Valera, Melanie F. Pradier, Maria Lomeli +1
Latent feature modeling allows capturing the latent structure responsible for generating the observed properties of a set of objects. It is often used to make predictions either fo…
Evaluation data contamination in LLMs: how do we measure it and (when) does it matter?
Aaditya K. Singh, Muhammed Yusuf Kocyigit, Andrew Poulton +4
Hampering the interpretation of benchmark scores, evaluation data contamination has become a growing concern in the evaluation of LLMs, and an active area of research studies its e…
TOOLVERIFIER: Generalization to New Tools via Self-Verification
Dheeraj Mekala, Jason Weston, Jack Lanchantin +4
Teaching language models to use tools is an important milestone towards building general assistants, but remains an open problem. While there has been significant progress on learn…
Toolformer: Language Models Can Teach Themselves to Use Tools
Timo Schick, Jane Dwivedi-Yu, Roberto Dessì +5
Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with…
General Bayesian inference schemes in infinite mixture models
Maria Lomeli
Bayesian statistical models allow us to formalise our knowledge about the world and reason about our uncertainty, but there is a need for better procedures to accurately encode its…
Inference-time sparse attention with asymmetric indexing
Pierre-Emmanuel Mazaré, Gergely Szilvasy, Maria Lomeli +4
Self-attention in transformer models is an incremental associative memory that maps key vectors to value vectors. One way to speed up self-attention is to employ GPU-compatible vec…
Antithetic and Monte Carlo kernel estimators for partial rankings
Maria Lomeli, Mark Rowland, Arthur Gretton +1
In the modern age, rankings data is ubiquitous and it is useful for a variety of applications such as recommender systems, multi-object tracking and preference learning. However, m…
Short window attention enables long-term memorization
Loïc Cabannes, Maximilian Beck, Gergely Szilvasy +6
Recent works show that hybrid architectures combining local sliding window attention layers and global attention layers outperform either of these architectures taken separately. H…
In-context Pretraining: Language Modeling Beyond Document Boundaries
Weijia Shi, Sewon Min, Maria Lomeli +9
Large language models (LMs) are currently trained to predict tokens given document prefixes, enabling them to directly perform long-form generation and prompting-style tasks which…
RA-DIT: Retrieval-Augmented Dual Instruction Tuning
Xi Victoria Lin, Xilun Chen, Mingda Chen +9
Retrieval-augmented language models (RALMs) improve performance by accessing long-tail and up-to-date knowledge from external data stores, but are challenging to build. Existing ap…
Source2Synth: Synthetic Data Generation and Curation Grounded in Real Data Sources
Alisia Lupidi, Carlos Gemmell, Nicola Cancedda +5
Synthetic data generation has recently emerged as a promising approach for enhancing the capabilities of large language models (LLMs) without the need for expensive human annotatio…
Self-Pruned Key-Value Attention: Learning When to Write by Predicting Future Utility
Gergely Szilvasy, Manuel Faysse, Maria Lomeli +5
Under modern test-time compute and agentic paradigms, language models process ever-longer sequences. Efficient text generation with transformer architectures is increasingly constr…
Universal Marginaliser for Deep Amortised Inference for Probabilistic Programs
Robert Walecki, Kostis Gourgoulias, Adam Baker +7
Probabilistic programming languages (PPLs) are powerful modelling tools which allow to formalise our knowledge about the world and reason about its inherent uncertainty. Inference…
A hybrid sampler for Poisson-Kingman mixture models
Maria Lomeli, Stefano Favaro, Yee Whye Teh
This paper concerns the introduction of a new Markov Chain Monte Carlo scheme for posterior sampling in Bayesian nonparametric mixture models with priors that belong to the general…
Augmented Language Models: a Survey
Grégoire Mialon, Roberto Dessì, Maria Lomeli +10
This survey reviews works in which language models (LMs) are augmented with reasoning skills and the ability to use tools. The former is defined as decomposing a potentially comple…
EditEval: An Instruction-Based Benchmark for Text Improvements
Jane Dwivedi-Yu, Timo Schick, Zhengbao Jiang +6
Evaluation of text generation to date has primarily focused on content created sequentially, rather than improvements on a piece of text. Writing, however, is naturally an iterativ…
The Faiss library
Matthijs Douze, Alexandr Guzhva, Chengqi Deng +6
Vector databases typically manage large collections of embedding vectors. Currently, AI applications are growing rapidly, and so is the number of embeddings that need to be stored…
Masking schemes for universal marginalisers
Divya Gautam, Maria Lomeli, Kostis Gourgoulias +2
We consider the effect of structure-agnostic and structure-dependent masking schemes when training a universal marginaliser (arXiv:1711.00695) in order to learn conditional distrib…
Sparse Delta Memory: Scaling the State of Linear RNNs through Sparsity
Loïc Cabannes, Pierre-Emmanuel Mazaré, Gergely Szilvasy +6
Linear attention models allow a fixed state size and a fixed amount of compute per token. However, due to their limited state size, linear attention models fall behind in long-cont…