Search Results for author:
Found 63 papers, 35 papers with code
Mixture-of-Depths: Dynamically allocating compute in transformer-based language models
Our method enforces a total compute budget by capping the number of tokens ($k$) that can participate in the self-attention and MLP computations at a given layer.
Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context
In this report, we present the latest model of the Gemini family, Gemini 1. 5 Pro, a highly compute-efficient multimodal mixture-of-experts model capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio.
Ranked #19 on
Code Generation
on HumanEval
Gemini: A Family of Highly Capable Multimodal Models
This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding.
Ranked #1 on
Multi-task Language Understanding
on MMLU
(using extra training data)
Android in the Wild: A Large-Scale Dataset for Android Device Control
The dataset contains human demonstrations of device interactions, including the screens and actions, and corresponding natural language instructions.
Mastering Diverse Domains through World Models
Developing a general algorithm that learns to solve tasks across a wide range of applications has been a fundamental challenge in artificial intelligence.
Ranked #6 on
Atari Games 100k
on Atari 100k
Improving Multimodal Interactive Agents with Reinforcement Learning from Human Feedback
Here we demonstrate how to use reinforcement learning from human feedback (RLHF) to improve upon simulated, embodied agents trained to a base level of competency with imitation learning.
Evaluating Long-Term Memory in 3D Mazes
However, most benchmark tasks in reinforcement learning do not test long-term memory in agents, slowing down progress in this important research direction.
Toward Next-Generation Artificial Intelligence: Catalyzing the NeuroAI Revolution
Neuroscience has long been an essential driver of progress in artificial intelligence (AI).
Large-Scale Retrieval for Reinforcement Learning
Effective decision making involves flexibly relating past experiences and relevant contextual information to a novel situation.
Intra-agent speech permits zero-shot task acquisition
Human language learners are exposed to a trickle of informative, context-sensitive language, but a flood of raw sensory data.
Evaluating Multimodal Interactive Agents
Creating agents that can interact naturally with humans is a common goal in artificial intelligence (AI) research.
Equilibrium Aggregation: Encoding Sets via Optimization
Processing sets or other unordered, potentially variable-sized inputs in neural networks is usually handled by aggregating a number of input tensors into a single representation.
Retrieval-Augmented Reinforcement Learning
In this paper we explore an alternative paradigm in which we train a network to map a dataset of past experiences to optimal behavior.
A data-driven approach for learning to control computers
It would be useful for machines to use computers as humans do so that they can aid us in everyday tasks.
Creating Multimodal Interactive Agents with Imitation and Self-Supervised Learning
A common vision from science fiction is that robots will one day inhabit our physical spaces, sense the world as we do, assist our physical labours, and communicate with us through natural language.
The functional specialization of visual cortex emerges from training parallel pathways with self-supervised predictive learning
We show that when we train a deep neural network architecture with two parallel pathways using a self-supervised predictive loss function, we can outperform other models in fitting mouse visual cortex.
Symbolic Behaviour in Artificial Intelligence
This approach will allow for AI to interpret something as symbolic on its own rather than simply manipulate things that are only symbols to human onlookers, and thus will ultimately lead to AI with more human-like symbolic fluency.
Imitating Interactive Intelligence
These evaluations convincingly demonstrate that interactive training and auxiliary losses improve agent behaviour beyond what is achieved by supervised learning of actions alone.
A meta-learning approach to (re)discover plasticity rules that carve a desired function into a neural network
Instead of experimental data, the rules are constrained by the functions they implement and the structure they are meant to produce.
Mastering Atari with Discrete World Models
The world model uses discrete representations and is trained separately from the policy.
Ranked #3 on
Atari Games
on Atari 2600 Skiing
(using extra training data)
Beyond Tabula-Rasa: a Modular Reinforcement Learning Approach for Physically Embedded 3D Sokoban
We explore whether integrated tasks like Mujoban can be solved by composing RL modules together in a sense-plan-act hierarchy, where modules have well-defined roles similarly to classic robot architectures.
Physically Embedded Planning Problems: New Challenges for Reinforcement Learning
To encourage progress towards this goal we introduce a set of physically embedded planning problems and make them publicly available.
dm_control: Software and Tasks for Continuous Control
The dm_control software package is a collection of Python libraries and task suites for reinforcement learning agents in an articulated-body simulation.
Automated curriculum generation through setter-solver interactions
We demonstrate the success of our approach in rich but sparsely rewarding 2D and 3D environments, where an agent is tasked to achieve a single goal selected from a set of possible goals that varies between episodes, and identify challenges for future work.
Dream to Control: Learning Behaviors by Latent Imagination
Learned world models summarize an agent's experience to facilitate learning complex behaviors.
LOGAN: Latent Optimisation for Generative Adversarial Networks
Training generative adversarial networks requires balancing of delicate adversarial dynamics.
Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
When evaluated on Go, chess and shogi, without any knowledge of the game rules, MuZero matched the superhuman performance of the AlphaZero algorithm that was supplied with the game rules.
Ranked #1 on
Atari Games
on Atari 2600 Alien
Deep Compressed Sensing
CS is flexible and data efficient, but its application has been restricted by the strong assumption of sparsity and costly reconstruction process.
Reliable Uncertainty Estimates in Deep Neural Networks using Noise Contrastive Priors
NCPs are compatible with any model that can output uncertainty estimates, are easy to scale, and yield reliable uncertainty estimates throughout training.
Deep reinforcement learning with relational inductive biases
We introduce an approach for augmenting model-free deep reinforcement learning agents with a mechanism for relational reasoning over structured representations, which improves performance, learning efficiency, generalization, and interpretability.
Is coding a relevant metaphor for building AI? A commentary on "Is coding a relevant metaphor for the brain?", by Romain Brette
Brette contends that the neural coding metaphor is an invalid basis for theories of what the brain does.
Deep Learning without Weight Transport
Current algorithms for deep learning probably cannot run in the brain because they rely on weight transport, where forward-path neurons transmit their synaptic weights to a feedback path, in a way that is likely impossible biologically.
Learning to Make Analogies by Contrasting Abstract Relational Structure
Here, we study how analogical reasoning can be induced in neural networks that learn to perceive and reason about raw visual data.
An investigation of model-free planning
The field of reinforcement learning (RL) is facing increasingly challenging domains with combinatorial complexity.
Learning Attractor Dynamics for Generative Memory
Based on the idea of memory writing as inference, as proposed in the Kanerva Machine, we show that a likelihood-based Lyapunov function emerges from maximising the variational lower-bound of a generative memory.
Learning Latent Dynamics for Planning from Pixels
Planning has been very successful for control tasks with known environment dynamics.
Ranked #2 on
Continuous Control
on DeepMind Walker Walk (Images)
Optimizing Agent Behavior over Long Time Scales by Transporting Value
Humans spend a remarkable fraction of waking life engaged in acts of "mental time travel".
Episodic Curiosity through Reachability
One solution to this problem is to allow the agent to create rewards for itself - thus making rewards dense and more suitable for learning.
Noise Contrastive Priors for Functional Uncertainty
NCPs are compatible with any model that can output uncertainty estimates, are easy to scale, and yield reliable uncertainty estimates throughout training.
Assessing the Scalability of Biologically-Motivated Deep Learning Algorithms and Architectures
Here we present results on scaling up biologically motivated models of deep learning on datasets which need deep networks with appropriate architectures to achieve good performance.
Measuring abstract reasoning in neural networks
To succeed at this challenge, models must cope with various generalisation `regimes' in which the training and test data differ in clearly-defined ways.
Relational recurrent neural networks
Memory-based neural networks model temporal data by leveraging an ability to remember information for long periods.
Ranked #72 on
Language Modelling
on WikiText-103
Relational Deep Reinforcement Learning
We introduce an approach for deep reinforcement learning (RL) that improves upon the efficiency, generalization capacity, and interpretability of conventional approaches through structured perception and relational reasoning.
Distributed Distributional Deterministic Policy Gradients
This work adopts the very successful distributional perspective on reinforcement learning and adapts it to the continuous control setting.
The Kanerva Machine: A Generative Distributed Memory
We present an end-to-end trained memory system that quickly adapts to new data and generates samples like them.
Recall Traces: Backtracking Models for Efficient Reinforcement Learning
In many environments only a tiny subset of all states yield high reward.
Unsupervised Predictive Memory in a Goal-Directed Agent
Animals execute goal-directed behaviours despite the limited range and scope of their sensors.
DeepMind Control Suite
The DeepMind Control Suite is a set of continuous control tasks with a standardised structure and interpretable rewards, intended to serve as performance benchmarks for reinforcement learning agents.
Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm
The game of chess is the most widely-studied domain in the history of artificial intelligence.
Ranked #1 on
Game of Go
on ELO Ratings
StarCraft II: A New Challenge for Reinforcement Learning
Finally, we present initial baseline results for canonical deep reinforcement learning agents applied to the StarCraft II domain.
Ranked #1 on
Starcraft II
on MoveToBeacon
A simple neural network module for relational reasoning
Relational reasoning is a central component of generally intelligent behavior, but has proven difficult for neural networks to learn.
Image Retrieval with Multi-Modal Query
Question Answering
+2
Interpolated Policy Gradient: Merging On-Policy and Off-Policy Gradient Estimation for Deep Reinforcement Learning
Off-policy model-free deep reinforcement learning methods using previously collected data can improve sample efficiency over on-policy policy gradient techniques.
Data-efficient Deep Reinforcement Learning for Dexterous Manipulation
Solving this difficult and practically relevant problem in the real world is an important long-term goal for the field of robotics.
Discovering objects and their relations from entangled scene representations
We show that RNs are capable of learning object relations from scene description data.
Generative Temporal Models with Memory
We consider the general problem of modeling temporal data with long-range dependencies, wherein new observations are fully or partially predictable based on temporally-distant, past observations.
Q-Prop: Sample-Efficient Policy Gradient with An Off-Policy Critic
We analyze the connection between Q-Prop and existing model-free algorithms, and use control variate theory to derive two variants of Q-Prop with conservative and aggressive adaptation.
Learning and Transfer of Modulated Locomotor Controllers
We study a novel architecture and training procedure for locomotion tasks.
Deep Reinforcement Learning for Robotic Manipulation with Asynchronous Off-Policy Updates
In this paper, we demonstrate that a recent deep reinforcement learning algorithm based on off-policy training of deep Q-functions can scale to complex 3D manipulation tasks and can learn deep neural network policies efficiently enough to train on real physical robots.
Matching Networks for One Shot Learning
Our algorithm improves one-shot accuracy on ImageNet from 87. 6% to 93. 2% and from 88. 0% to 93. 8% on Omniglot compared to competing approaches.
One-shot Learning with Memory-Augmented Neural Networks
Despite recent breakthroughs in the applications of deep neural networks, one setting that presents a persistent challenge is that of "one-shot learning."
Continuous Deep Q-Learning with Model-based Acceleration
In this paper, we explore algorithms and representations to reduce the sample complexity of deep reinforcement learning for continuous control tasks.
Deep Reinforcement Learning in Large Discrete Action Spaces
Being able to reason in an environment with a large number of discrete actions is essential to bringing reinforcement learning to a larger class of problems.
Learning Continuous Control Policies by Stochastic Value Gradients
One of these variants, SVG(1), shows the effectiveness of learning models, value functions, and policies simultaneously in continuous domains.
