arXiv subjects

Piotr Mirowski

Publications and source records attributed to Piotr Mirowski.

22 records · Page 2Linked to original sources

arXiv2017

Improvised Comedy as a Turing Test

The best improvisational theatre actors can make any scene partner, of any skill level or ability, appear talented and proficient in the art form, and thus "make them shine". To challenge this improvisational paradigm, we built an artificial intelligence (AI) trained to perform live shows alongside human actors for human audiences. Over the course of 30 performances to a combined audience of almost 3000 people, we have refined theatrical games which involve combinations of human and (at times, adversarial) AI actors. We have developed specific scene structures to include audience participants in interesting ways. Finally, we developed a complete show structure that submitted the audience to a Turing test and observed their suspension of disbelief, which we believe is key for human/non-human theatre co-creation.

cs.AI↗

arXiv2017

One-Shot Reinforcement Learning for Robot Navigation with Interactive Replay

Recently, model-free reinforcement learning algorithms have been shown to solve challenging problems by learning from extensive interaction with the environment. A significant issue with transferring this success to the robotics domain is that interaction with the real world is costly, but training on limited experience is prone to overfitting. We present a method for learning to navigate, to a fixed goal and in a known environment, on a mobile robot. The robot leverages an interactive world model built from a single traversal of the environment, a pre-trained visual feature encoder, and stochastic environmental augmentation, to demonstrate successful zero-shot transfer under real-world environmental variations without fine-tuning.

cs.AI↗

arXiv2017

Learning to Navigate in Complex Environments

Learning to navigate in complex environments with dynamic elements is an important milestone in developing AI agents. In this work we formulate the navigation question as a reinforcement learning problem and show that data efficiency and task performance can be dramatically improved by relying on additional auxiliary tasks leveraging multimodal sensory inputs. In particular we consider jointly learning the goal-driven reinforcement learning problem with auxiliary depth prediction and loop closure classification tasks. This approach can learn to navigate from raw sensory input in complicated 3D mazes, approaching human-level performance even under conditions where the goal location changes frequently. We provide detailed analysis of the agent behaviour, its ability to localise, and its network activity dynamics, showing that the agent implicitly learns key navigation abilities.

cs.AI↗

arXiv2015

Dependency Recurrent Neural Language Models for Sentence Completion

Recent work on language modelling has shifted focus from count-based models to neural models. In these works, the words in each sentence are always considered in a left-to-right order. In this paper we show how we can improve the performance of the recurrent neural network (RNN) language model by incorporating the syntactic dependencies of a sentence, which have the effect of bringing relevant contexts closer to the word being predicted. We evaluate our approach on the Microsoft Research Sentence Completion Challenge and show that the dependency RNN proposed improves over the RNN by about 10 points in accuracy. Furthermore, we achieve results comparable with the state-of-the-art models on this task.

cs.CL↗