Discover DeepMind’s cutting-edge research presented at ICLR 2023

Introduction:

DeepMind researchers are set to present their latest AI advancements at the 11th International Conference on Learning Representations in Kigali, Rwanda. With 23 papers covering topics from AGI to AI for science, the conference promises to be a melting pot of innovative ideas and approaches that could revolutionize the field. Learn more about the conference and the exciting developments that will be showcased. Find out more at DeepMind events/events/iclr-2023/resources.

Full News:

Research towards AI models that can generalise, scale, and accelerate science

Next week marks the start of the 11th International Conference on Learning Representations (ICLR), taking place 1-5 May in Kigali, Rwanda. This will be the first major artificial intelligence (AI) conference to be hosted in Africa and the first in-person event since the start of the pandemic.

Researchers from around the world will gather to share their cutting-edge work in deep learning spanning the fields of AI, statistics and data science, and applications including machine vision, gaming and robotics. We’re proud to support the conference as a Diamond sponsor and DEI champion.

Teams from across DeepMind are presenting 23 papers this year. Here are a few highlights:

Open questions on the path to AGI

Recent progress has shown AI’s incredible performance in text and image, but more research is needed for systems to generalise across domains and scales. This will be a crucial step on the path to developing artificial general intelligence (AGI) as a transformative tool in our everyday lives.

We present a new approach where models learn by solving two problems in one. By training models to look at a problem from two perspectives at the same time, they learn how to reason on tasks that require solving similar problems, which is beneficial for generalisation. We also explored the capability of neural networks to generalise by comparing them to the Chomsky hierarchy of languages. By rigorously testing 2200 models across 16 different tasks, we uncovered that certain models struggle to generalise, and found that augmenting them with external memory is crucial to improve performance.

Another challenge we tackle is how to make progress on longer-term tasks at an expert-level, where rewards are few and far between. We developed a new approach and open-source training data set to help models learn to explore in human-like ways over long time horizons.

Innovative approaches

As we develop more advanced AI capabilities, we must ensure current methods work as intended and efficiently for the real world. For example, although language models can produce impressive answers, many cannot explain their responses. We introduce a method for using language models to solve multi-step reasoning problems by exploiting their underlying logical structure, providing explanations that can be understood and checked by humans. On the other hand, adversarial attacks are a way of probing the limits of AI models by pushing them to create wrong or harmful outputs. Training on adversarial examples makes models more robust to attacks, but can come at the cost of performance on ‘regular’ inputs. We show that by adding adapters, we can create models that allow us to control this tradeoff on the fly.

Reinforcement learning (RL) has proved successful for a range of real-world challenges, but RL algorithms are usually designed to do one task well and struggle to generalise to new ones. We propose algorithm distillation, a method that enables a single model to efficiently generalise to new tasks by training a transformer to imitate the learning histories of RL algorithms across diverse tasks. RL models also learn by trial and error which can be very data-intensive and time-consuming. It took nearly 80 billion frames of data for our model Agent 57 to reach human-level performance across 57 Atari games. We share a new way to train to this level using 200 times less experience, vastly reducing computing and energy costs.

AI for science

AI is a powerful tool for researchers to analyse vast amounts of complex data and understand the world around us. Several papers show how AI is accelerating scientific progress – and how science is advancing AI.

Predicting a molecule’s properties from its 3D structure is critical for drug discovery. We present a denoising method that achieves a new state-of-the-art in molecular property prediction, allows large-scale pre-training, and generalises across different biological datasets. We also introduce a new transformer which can make more accurate quantum chemistry calculations using data on atomic positions alone.

Finally, with FIGnet, we draw inspiration from physics to model collisions between complex shapes, like a teapot or a doughnut. This simulator could have applications across robotics, graphics and mechanical design.

See the full list of DeepMind papers and schedule of events at ICLR 2023.

Conclusion:

In conclusion, the 11th International Conference on Learning Representations will take place in Kigali, Rwanda from May 1-5. AI researchers from around the world will present their work, covering topics such as generalizing AI models, innovative approaches in AI, and the use of AI in scientific research. DeepMind is proud to support the conference and present 23 papers on topics such as open questions in the path to AGI, innovative approaches in AI, and AI for science. These papers aim to push the boundaries of AI research and accelerate scientific progress. For more details, visit the DeepMind website.

Frequently Asked Questions:

## What is DeepMind’s latest research at ICLR 2023?

DeepMind’s latest research at ICLR 2023 focuses on advancements in AI and machine learning algorithms, particularly in the fields of reinforcement learning, computer vision, and natural language processing.

## What are some key highlights from DeepMind’s research at ICLR 2023?

Some key highlights from DeepMind’s research at ICLR 2023 include breakthroughs in self-supervised learning, improved sample efficiency in reinforcement learning, and advancements in multimodal AI models that can reason across multiple modalities.

## How does DeepMind’s research at ICLR 2023 contribute to the field of AI and machine learning?

DeepMind’s research at ICLR 2023 contributes to the field of AI and machine learning by pushing the boundaries of what is possible with current algorithms, leading to more efficient and intelligent AI systems that can tackle complex real-world problems.

## What are some potential applications of DeepMind’s latest research at ICLR 2023?

Some potential applications of DeepMind’s latest research at ICLR 2023 include more capable autonomous agents, improved medical image analysis and diagnosis, and more advanced natural language understanding systems.

## How does DeepMind’s research at ICLR 2023 advance the state of the art in reinforcement learning?

DeepMind’s research at ICLR 2023 advances the state of the art in reinforcement learning by developing novel algorithms that improve sample efficiency, generalization, and robustness of reinforcement learning agents.

## What are the implications of DeepMind’s research at ICLR 2023 for the field of computer vision?

The implications of DeepMind’s research at ICLR 2023 for the field of computer vision include more accurate and efficient object recognition, improved scene understanding, and advancements in visual reasoning tasks.

## How does DeepMind’s research at ICLR 2023 impact natural language processing (NLP) tasks?

DeepMind’s research at ICLR 2023 impacts natural language processing (NLP) tasks by addressing challenges such as language understanding and generation, enabling more sophisticated and context-aware NLP systems.

## What sets DeepMind’s research at ICLR 2023 apart from previous work in the field?

DeepMind’s research at ICLR 2023 sets itself apart from previous work in the field by introducing novel methodologies, achieving state-of-the-art results, and addressing some of the most challenging problems in AI and machine learning.

## How will DeepMind’s latest research at ICLR 2023 benefit the broader scientific community?

DeepMind’s latest research at ICLR 2023 will benefit the broader scientific community by sharing new insights, methodologies, and findings that can be leveraged by other researchers and organizations to advance the state of the art in AI and machine learning.

## What can we expect from DeepMind in the future based on their latest research at ICLR 2023?

Based on their latest research at ICLR 2023, we can expect DeepMind to continue to push the boundaries of AI and machine learning, leading to more impactful and far-reaching advancements in the field in the years to come.