Announcing Raia Hadsell (Google DeepMind) at RAAIS 2026

2025-10-09 · Source: Air Street Press · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Expert, short

Summary

Raia Hadsell, VP of Research at Google DeepMind and co-lead of its Frontier AI unit, is announced as a returning speaker for the 10th annual Research and Applied AI Summit (RAAIS) on June 12th, 2026, in London. Her career, spanning from a Senior Research Scientist at DeepMind in 2017 to her current role, consistently focuses on challenging open problems like continual and transfer learning, deep reinforcement learning for robotics, and navigation. Hadsell's foundational work includes shaping metric learning and Siamese neural networks, combating catastrophic forgetting with "elastic weight consolidation," and contributing to generalist robotic agents like RoboCat and bipedal robot locomotion. Her recent contributions extend to frontier language models such as Gemini 2.5, Gemma 2, and RecurrentGemma. Beyond research, she founded "Transactions on Machine Learning Research" (TMLR) and serves as a UK government AI Ambassador, chairing national AI research initiatives.

Key takeaway

For AI Scientists and Research Scientists developing robust, adaptive AI systems, Raia Hadsell's career highlights continual learning and embodied intelligence. You should prioritize research into methods like elastic weight consolidation. This mitigates catastrophic forgetting, allowing models to improve over time without full retraining. Consider her work on generalist robotic agents and frontier language models. This provides a blueprint for integrating foundational research into practical, real-world AI applications.

Key insights

The article highlights Raia Hadsell's consistent focus on lifelong learning and embodied AI, bridging foundational research with frontier systems.

Principles

• Catastrophic forgetting is a deep obstacle for AI systems.
• Lifelong learning requires protecting learned parameters.
• Embodied AI benefits from continual adaptation.

Method

Elastic weight consolidation protects important neural network parameters while acquiring new knowledge to overcome catastrophic forgetting.

In practice

• Apply elastic weight consolidation for continual learning.
• Explore Siamese neural networks for representation learning.
• Investigate RoboCat for robotic manipulation tasks.

Topics

• Continual Learning
• Deep Reinforcement Learning
• Embodied AI
• Frontier AI Systems
• Robotic Agents
• AI Policy

Best for: AI Scientist, Research Scientist, Director of AI/ML

Related on AIssential

• PHASER: Phase-Aware and Semantic Experience Replay for Vision-Language-Action Models — PHASER prevents catastrophic forgetting in VLA models by intelligently prioritizing critical sub-skills and high-risk phases during experience replay.
• Google's "Infinite Learning" and OpenAI's leaked "AI Pen" — Continual learning, inspired by human neuroplasticity, is critical for LLMs to acquire and retain knowledge over time.
• Forgetting is Not Erasure: Recovering Latent Knowledge via Transport Keys — Catastrophic forgetting often reflects interface drift, not erasure, allowing latent knowledge recovery via alignment.
• Octopus: History-Free Gradient Orthogonalization for Continual Learning in Multimodal Large Language Models — Octopus enables continual learning in MLLMs by orthogonalizing gradients without needing historical data.
• Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments — Continual learning benchmarks must isolate online improvement from static model capabilities using expert-validated, latent-structure tasks.
• Bounding the Predictive Space: How Topological AI Solves Catastrophic Forgetting Through… — Topological AI uses prime-indexed embedding anchors to prevent catastrophic forgetting in LLMs while retaining cognitive flexibility.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Air Street Press.