Post

How to get URL link on X (Twitter) App

On the Twitter thread, click on or icon on the bottom
Click again on or Share Via icon
Click on Copy Link to Tweet
Paste it above and click "Unroll Thread"!
More info at Twitter Help

Google DeepMind

Sep 9, 2021 • 10 tweets • 9 min read • Read on X

Scrolly

@ai_ucl

Introducing the '21 DeepMind x @ai_ucl Reinforcement Learning Lecture Series, a comprehensive introduction to modern RL.

Follow along with our researchers are they explore Markov Decision Processes, sample-based learning algorithms & much more: dpmd.ai/2021RLseries 1/2

@YouTube

Also find the full series via the DeepMind @YouTube channel: dpmd.ai/DeepMindxUCL21

@ai_ucl

In the first lecture of the series, Research Scientist Hado introduces the course and explores the fascinating connection between reinforcement learning and artificial intelligence: dpmd.ai/RLseries1

#DeepMindxUCL @ai_ucl

@ai_ucl

In lecture two, Research Scientist Hado explains why it's important for learning agents to balance exploring and exploiting acquired knowledge at the same time: dpmd.ai/RLseries2

#DeepMindxUCL @ai_ucl

@ai_ucl

In the third lecture, Research Scientist Diana shows us how to solve MDPs with dynamic programming to extract accurate predictions and good control policies: dpmd.ai/RLseries3

#DeepMindxUCL @ai_ucl

@ai_ucl

In lecture four, Diana covers dynamic programming algorithms as contraction mappings, looking at when and how they converge to the right solutions: dpmd.ai/RLseries4

#DeepMindxUCL @ai_ucl

@ai_ucl

In this lecture, Hado explores model-free prediction and its relation to Monte Carlo and temporal difference algorithms: dpmd.ai/RLseries5

#DeepMindxUCL @ai_ucl

@ai_ucl

In part two of the model-free lecture, Hado explains how to use prediction algorithms for policy improvement, leading to algorithms - like Q-learning - that can learn good behaviour policies from sampled experience: dpmd.ai/RLseries6

#DeepMindxUCL @ai_ucl

@ai_ucl

In this lecture, Hado explains how to combine deep learning with reinforcement learning for deep reinforcement learning. He looks at the properties and difficulties that arise when combining function approximation with RL algorithms: dpmd.ai/RLseries7

#DeepMindxUCL @ai_ucl

@ai_ucl

In this lecture, Research Engineer Matteo explains how to learn and use models, including algorithms like Dyna and Monte-Carlo tree search (MCTS): dpmd.ai/RLseries8

#DeepMindxUCL @ai_ucl

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @GoogleDeepMind

Google DeepMind

@GoogleDeepMind

Jul 23

Our new state-of-the-art AI model Aeneas transforms how historians connect the past. 📜

Ancient inscriptions often lack context – it's like solving a puzzle with 90% of the pieces lost to time. It helps researchers interpret and situate inscriptions in their past context. 🧵

By transforming each ancient text into a unique historical fingerprint, Aeneas can identify similarities across 176,000 Latin inscriptions.

In our study, historians found these ‘parallels’ to be helpful research starting points 9 out of 10 times - improving their confidence by 44%.

We tested Aeneas on the Res Gestae Divi Augusti – one of the most debated inscriptions.

Without prior knowledge, it successfully mapped out the leading scholarly theories on its dating, showing how AI can help model history in a quantitative way. 📊

Read 4 tweets

Google DeepMind

@GoogleDeepMind

Jun 24

We’re bringing powerful AI directly onto robots with Gemini Robotics On-Device. 🤖

It’s our first vision-language-action model to help make robots faster, highly efficient, and adaptable to new tasks and environments - without needing a constant internet connection. 🧵

What makes this new model unique?

🔵 It has the generality and dexterity of Gemini Robotics - but it can run locally on the device
🔵 It can handle a wide variety of complex, two-handed tasks out of the box
🔵 It can learn new skills with as few as 50-100 demonstrations

From humanoids to industrial bi-arm robots, the model supports multiple embodiments, even though it was pre-trained on ALOHA - while following instructions from humans. 💬

These tasks may seem easy for us but require fine motor skills, precise manipulation and more. ↓

Read 5 tweets

Google DeepMind

@GoogleDeepMind

Jun 17

Hot Gemini updates off the press. 🚀

Anyone can now use 2.5 Flash and Pro to build and scale production-ready AI applications. 🙌

We’re also launching 2.5 Flash-Lite in preview: the fastest model in the 2.5 family to respond to requests, with the lowest cost too. 🧵

2.5 Flash-Lite now supports:

🔹Thinking: improving performance and transparency through step-by-step reasoning
🔹Tool-use: including Search, code execution and 1 million token context window - similar to 2.5 Flash and Pro

⚡ 2.5 Flash-Lite is our most cost efficient model yet - and with lower latency than 2.0 Flash-Lite and Flash on a broad sample of prompts.

It also has all-around, higher quality than 2.0 Flash-Lite on coding, math, science, reasoning and multimodal benchmarks.

Read 5 tweets

Google DeepMind

@GoogleDeepMind

May 14

Introducing AlphaEvolve: a Gemini-powered coding agent for algorithm discovery.

It’s able to:

🔘 Design faster matrix multiplication algorithms
🔘 Find new solutions to open math problems
🔘 Make data centers, chip design and AI training more efficient across @Google. 🧵

Our system uses:
🔵 LLMs: To synthesize information about problems as well as previous attempts to solve them - and to propose new versions of algorithms
🔵 Automated evaluation: To address the broad class of problems where progress can be clearly and systematically measured.
🔵 Evolution: Iteratively improving the best algorithms found, and re-combining ideas from different solutions to find even better ones.

Over the past year, we’ve deployed algorithms discovered by AlphaEvolve across @Google’s computing ecosystem, including data centers, software and hardware.

It’s been able to:

🔧 Optimize data center scheduling
🔧 Assist in hardware design
🔧 Enhance AI training and inference

Read 6 tweets

Google DeepMind

@GoogleDeepMind

Apr 30

We’re helping robots self-improve with the power of LLMs. 🤖

Introducing the Summarize, Analyze, Synthesize (SAS) prompt, which analyzes how they perform tasks based on previous actions and then suggests ways for them to get better using the medium of table tennis. 🏓

Large language models like Gemini have an inherent ability to problem solve, without needing to retrain for specific jobs.

Robots can use these models to improve how they operate over time, by interacting with the world, and learning from those interactions. 🦾 goo.gle/4jVFsoE

With the SAS prompt, we can now use language models like Gemini to learn from a robot's history.

This allows the model to analyze parameter effects ⚡, and suggest ways to improve - similar to a real-life table tennis coach. 💡 goo.gle/3GvWQ54

Read 4 tweets

Google DeepMind

@GoogleDeepMind

Apr 23

We built an AI model to simulate how a fruit fly walks, flies and behaves – in partnership with @HHMIJanelia. 🪰

Our computerized insect replicates realistic motion, and can even use its eyes to control its actions.

Here’s how we developed it – and what it means for science. 🧵

To create it, we turned to MuJoCo, our open-source physics simulator – created for robotics and biomechanics – and added features such as:

▪️simulating fluid forces on the flapping wings, enabling flight
▪️adhesion actuators – mimicking the gripping force of insect feet

We then trained an artificial neural network on real fly behavior from recorded videos 🎥 and let it control the model in MuJoCo.

This enables it to learn how to move the virtual insect in the most realistic way.