Building interactive agents in video game worlds

Building interactive agents in video game worlds

Most artificial intelligence (AI) researchers now believe that writing computer code which can capture the nuances of situated interactions is impossible. Alternatively, modern machine learning (ML) researchers have focused on learning about these types of interactions from data. To explore these learning-based approaches and quickly build agents that can make sense of human instructions and safely perform actions in open-ended conditions, we created a research framework within a video game environment.Today, we’re publishing a paper [INSERT LINK] and collection of videos, showing our early steps in building video game AIs that can understand fuzzy human concepts – and therefore, can begin to interact with people on their own terms.Read More

Benchmarking the next generation of never-ending learners

Our new paper, NEVIS’22: A Stream of 100 Tasks Sampled From 30 Years of Computer Vision Research, proposes a playground to study the question of efficient knowledge transfer in a controlled and reproducible setting. The Never-Ending Visual classification Stream (NEVIS’22) is a benchmark stream in addition to an evaluation protocol, a set of initial baselines, and an open-source codebase. This package provides an opportunity for researchers to explore how models can continually build on their knowledge to learn future tasks more efficiently.Read More

Best practices for data enrichment

Best practices for data enrichment

At DeepMind, our goal is to make sure everything we do meets the highest standards of safety and ethics, in line with our Operating Principles. One of the most important places this starts with is how we collect our data. In the past 12 months, we’ve collaborated with Partnership on AI (PAI) to carefully consider these challenges, and have co-developed standardised best practices and processes for responsible human data collection.Read More

On the Expressivity of Markov Reward

Our main results prove that while reward can express many tasks, there exist instances of each task type that no Markov reward function can capture. We then provide a set of polynomial-time algorithms that construct a reward function which allows an agent to optimize tasks of each of these three types, and correctly determine when no such reward function exists.Read More

Measuring perception in AI models

Perception – the process of experiencing the world through senses – is a significant part of intelligence. And building agents with human-level perceptual understanding of the world is a central but challenging task, which is becoming increasingly important in robotics, self-driving cars, personal assistants, medical imaging, and more. So today, we’re introducing the Perception Test, a multimodal benchmark using real-world videos to help evaluate the perception capabilities of a model.Read More

How undesired goals can arise with correct rewards

As we build increasingly advanced artificial intelligence (AI) systems, we want to make sure they don’t pursue undesired goals. Such behaviour in an AI agent is often the result of specification gaming – exploiting a poor choice of what they are rewarded for. In our latest paper, we explore a more subtle mechanism by which AI systems may unintentionally learn to pursue undesired goals: goal misgeneralisation (GMG). GMG occurs when a system’s capabilities generalise successfully but its goal does not generalise as desired, so the system competently pursues the wrong goal. Crucially, in contrast to specification gaming, GMG can occur even when the AI system is trained with a correct specification.Read More

Discovering novel algorithms with AlphaTensor

In our paper, published today in Nature, we introduce AlphaTensor, the first artificial intelligence (AI) system for discovering novel, efficient, and provably correct algorithms for fundamental tasks such as matrix multiplication. This sheds light on a 50-year-old open question in mathematics about finding the fastest way to multiply two matrices. This paper is a stepping stone in DeepMind’s mission to advance science and unlock the most fundamental problems using AI. Our system, AlphaTensor, builds upon AlphaZero, an agent that has shown superhuman performance on board games, like chess, Go and shogi, and this work shows the journey of AlphaZero from playing games to tackling unsolved mathematical problems for the first time. Read More