697 citations · 2.5k across the 36 of their papers we have counts for
14 papers · 1 filter
Vision-Language Model Dialog Games for Self-Improvement
Ksenia Konyushkova, Christos Kaplanis, Serkan Cabi +1
The increasing demand for high-quality, diverse training data poses a significant bottleneck in advancing vision-language models (VLMs). This paper presents VLM Dialog Games, a nov…
: Policy Representations with Successor Features
Gianluca Scarpellini, Ksenia Konyushkova, Claudio Fantacci +3
This paper describes , a method for representing behaviors of black box policies as feature vectors. The policy representations capture how the statistics of founda…
Active Offline Policy Selection
Ksenia Konyushkova, Yutian Chen, Tom Le Paine +5
This paper addresses the problem of policy selection in domains with abundant logged data, but with a restricted interaction budget. Solving this problem would enable safe evaluati…
Offline Learning from Demonstrations and Unlabeled Experience
Konrad Zolna, Alexander Novikov, Ksenia Konyushkova +6
Behavior cloning (BC) is often practical for robot learning because it allows a policy to be trained offline without rewards, by supervised learning on expert demonstrations. Howev…
Positive-Unlabeled Reward Learning
Danfei Xu, Misha Denil
Learning reward functions from data is a promising path towards achieving scalable Reinforcement Learning (RL) for robotics. However, a major challenge in training agents from lear…
Task-Relevant Adversarial Imitation Learning
Konrad Zolna, Scott Reed, Alexander Novikov +6
We show that a critical vulnerability in adversarial imitation is the tendency of discriminator networks to learn spurious associations between visual features and expert labels. W…
Making Efficient Use of Demonstrations to Solve Hard Exploration Problems
Tom Le Paine, Caglar Gulcehre, Bobak Shahriari +11
This paper introduces R2D3, an agent that makes efficient use of demonstrations to solve hard exploration problems in partially observable environments with highly variable initial…
Learned Optimizers that Scale and Generalize
Olga Wichrowska, Niru Maheswaranathan, Matthew W. Hoffman +4
Learning to learn has emerged as an important direction for achieving artificial intelligence. Two of the primary barriers to its adoption are an inability to scale to larger probl…
Noisy Activation Functions
Caglar Gulcehre, Marcin Moczulski, Misha Denil +1
Common nonlinear activation functions used in neural networks can cause training difficulties due to the saturation behavior of the activation function, which may hide dependencies…
ACDC: A Structured Efficient Linear Layer
Marcin Moczulski, Misha Denil, Jeremy Appleyard +1
The linear layer is one of the most pervasive modules in deep learning representations. However, it requires parameters and operations. These costs can be prohibi…
Deep Multi-Instance Transfer Learning
Dimitrios Kotzias, Misha Denil, Phil Blunsom +1
We present a new approach for transferring knowledge from groups to individuals that comprise them. We evaluate our method in text, by inferring the ratings of individual sentences…
Deep Fried Convnets
Zichao Yang, Marcin Moczulski, Misha Denil +4
The fully connected layers of a deep convolutional neural network typically contain over 90% of the network parameters, and consume the majority of the memory required to store the…
Predicting Parameters in Deep Learning
Misha Denil, Babak Shakibi, Laurent Dinh +2
We demonstrate that there is significant redundancy in the parameterization of several deep learning models. Given only a few weight values for each feature it is possible to accur…
Recklessly Approximate Sparse Coding
Misha Denil, Nando de Freitas
It has recently been observed that certain extremely simple feature encoding techniques are able to achieve state of the art performance on several standard image classification be…