A new model and dataset for long-range memory  00h00 10 février
This blog introduces a new long-range memory model, the Compressive Transformer, alongside a new benchmark for book-level language modelling, PG19. We provide the conceptual tools needed to understand this new research in the context of recent developments in memory models and language modelling.
AlphaFold: Using AI for scientific discovery  00h00 15 janvier
Our Nature paper describes AlphaFold, a system that generates3D models of proteins that are far more accurate than any that have come before.
Dopamine and temporal difference learning: A fruitful relationship between neuroscience and AI  00h00 15 janvier
A recent development in computer science may provide a deep, parsimonious explanation for several previously unexplained features of reward learning in the brain.
Using WaveNet technology to reunite speech-impaired users with their original voices  00h00 18 décembre
We demonstrate an early proof of concept of how text-to-speech technologies can synthesise a high-quality, natural sounding voice using minimal recorded speech data.
Learning human objectives by evaluating hypothetical behaviours  00h00 13 décembre
We present a new method for training reinforcement learning agents from human feedback in the presence of unknown unsafe states.
From unlikely start-up to major scientific organisation: Entering our tenth year at DeepMind  00h00 05 décembre
Weve come a long way in building the organisation we need to achieve our long-term mission.
Strengthening the AI community  00h00 21 novembre
AI requires people with different experiences, knowledge and backgrounds, which is why we started the DeepMind Scholarship programme and supportuniversitiesand the wider ecosystem.
Advanced machine learning helps Play Store users discover personalised apps  00h00 18 novembre
In collaboration with Google Play,our team that leads on collaborations with Googlehas driven significant improvements in the Play Store’s discovery systems, helping to deliver a more personalised and intuitive Play Store experience for users.
AlphaStar: Grandmaster level in StarCraft II using multi-agent reinforcement learning  00h00 30 octobre
AlphaStar is the first AI to reach the top league of a widely popular esport without any game restrictions.
Causal Bayesian Networks: A flexible tool to enable fairer machine learning  00h00 03 octobre
Decisions based on machine learning (ML) are potentially advantageous over human decisions, but the data used to train them often contains human and societal biases that can lead to harmful decisions.
DeepMind’s health team joins Google Health  00h00 18 septembre
Heres what the future looks like for the team.
Episode 8: Demis Hassabis - The interview  00h00 17 septembre
In this special extended episode, Hannah meets Demis Hassabis, the CEO and co-founder of DeepMind.
Episode 7: Towards the future  00h00 10 septembre
AI researchers around the world are trying to create a general purpose learning system that can learn to solve a broad range of problems without being taught how. Hannah explores the journey to get there.
Replay in biological and artificial neural networks  00h00 06 septembre
Our waking and sleeping lives are punctuated by fragments of recalled memories: a sudden connection in the shower between seemingly disparate thoughts, or an ill-fated choice decades ago that haunts us as we struggle to fall asleep.
Episode 6: AI for everyone  00h00 03 septembre
Hannah investigates the more human side of the technology, some ethical issues around how it is developed and used, and the efforts to create a future of AI that works for everyone.