The Curious Engineer

All blogs / AI / ML

Google DeepMind News

deepmind.google/blog

Discover our latest AI breakthroughs, projects, and updates.

AI / MLScience

’16’17’18’19’20’21’22’23’24’25’26
Posts
407 from 24 Feb 2016 to 28 Oct 2026
Posting
weekly (104 in the last year)
Kind
company blog
Feed
https://deepmind.google/blog/rss.xml

2026 86

  1. Gemini 4 Carbon: Our next frontier model
  2. Genie 4: A new generation of world simulation
  3. Beyond tree cover: mapping the world’s forest types
  4. Accelerating the search for antimicrobial peptides with Google Antigravity
  5. Google DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detection
  6. Developing Gemini 4 Argon responsibly and safely
  7. Gemini 4 Argon for Robotics
  8. Gemini 4 Argon: Built to solve your hardest problems
  9. EmbeddingGemma 2: an open, lightweight multimodal embedding model
  10. Gemini 4 Argon: our next era of frontier intelligence
  11. Introducing SynthID Bio
  12. Introducing Gemini 3.8 Live with Live Avatar
  13. Advancing Private AI Compute with secure, server-side memory
  14. Gemini 3.8 text-to-speech says hello
  15. Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
  16. AlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genome
  17. Introducing WeatherNext 3, our most advanced and accurate global weather AI model
  18. Proactive cyber defense for governments and enterprises
  19. Introducing Gemini 3.8 Flash and 3.8 Flash Cyber
  20. Introducing agentic video understanding with Gemini
  21. Gemini Omni 1.1 Flash lets you build with more control
  22. Piloting the world's first double-blind AI evaluations
  23. Intelligent transcription with Gemini 3.5 Transcribe
  24. From Atari to EVE Online: Building on 15 Years of AI Research in Games
  25. Introducing Gemini 3.7 Flash
  26. Putting sign language AI into users’ hands
  27. WeatherNext: AI model achieves breakthrough in forecasting cyclones
  28. Gemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration
  29. We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
  30. Gemini Robotics 2 brings whole body intelligence to robots
  31. Accelerating the frontiers of scientific discovery: Google’s $40M commitment to the Genesis Mission
  32. Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
  33. Introducing Gemini 3.5 Flash Cyber
  34. Our approach to bioresilience
  35. Empowering India’s next generation of innovators with ATL Saathi
  36. Google DeepMind and A24 announce first-of-its-kind research partnership
  37. Start building with Nano Banana 2 Lite and Gemini Omni Flash
  38. Introducing computer use in Gemini 3.5 Flash
  39. Unlocking UK house-building with AI-accelerated planning
  40. Securing the future of AI agents
  41. DiffusionGemma: 4x faster text generation
  42. Investing in multi-agent AI safety research
  43. Fluid, natural voice translation with Gemini 3.5 Live Translate
  44. Introducing Gemma 4 12B: a unified, encoder-free multimodal model
  45. Powering the future of robotics in Europe
  46. Measuring the impact of learning with AI in Sierra Leone and beyond
  47. We’re launching the Google DeepMind Accelerator program in Asia Pacific to tackle environmental risks
  48. Fast-tracking genetic leads to reverse cellular aging
  49. Simulate real-world places with Project Genie and Street View
  50. Introducing Gemini Omni
  51. Introducing Google Antigravity 2.0
  52. Gemini for Science: AI experiments and tools for a new era of discovery
  53. Making it easier to understand how content was created and edited
  54. Strengthening Singapore’s AI Future: A New National Partnership
  55. Finding the molecular switches behind new infectious diseases
  56. Opening new paths in aging research
  57. Accelerating discovery of liver disease mechanisms
  58. Uniting biological toolkits for a new approach to ALS
  59. Uncovering repurposed medicines to fight liver fibrosis
  60. How WeatherNext helped the National Hurricane Center better predict Hurricane Melissa’s historic landfall in Jamaica
  61. Gemini 3.5: frontier intelligence with action
  62. Co-Scientist: A multi-agent AI partner to accelerate research
  63. AlphaEvolve: How our Gemini-powered coding agent is scaling impact across fields
  64. Enabling a new model for healthcare with AI co-clinician
  65. Announcing our partnership with the Republic of Korea
  66. Decoupled DiLoCo: A new frontier for resilient, distributed AI training
  67. Partnering with industry leaders to accelerate AI transformation
  68. Gemini 3.1 Flash TTS: the next generation of expressive AI speech
  69. Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning
  70. Gemma 4: Byte for byte, the most capable open models
  71. Reimagining the mouse pointer for the AI era
  72. Gemini 3.1 Flash Live: Making audio AI more natural and reliable
  73. Protecting people from harmful manipulation
  74. Lyria 3 Pro: Create longer tracks in more
  75. Measuring progress toward AGI: A cognitive framework
  76. From games to biology and beyond: 10 years of AlphaGo’s impact
  77. Gemini 3.1 Flash-Lite: Built for intelligence at scale
  78. Nano Banana 2: Combining Pro capabilities with lightning-fast speed
  79. Gemini 3.1 Pro: A smarter model for your most complex tasks
  80. A new way to express yourself: Gemini can now create music
  81. Accelerating discovery in India through AI-powered science and education
  82. Gemini 3 Deep Think: Advancing science, research and engineering
  83. Accelerating Mathematical and Scientific Discovery with Gemini Deep Think
  84. Project Genie: Experimenting with infinite, interactive worlds
  85. D4RT: Teaching AI to see the world in four dimensions
  86. Veo 3.1 Ingredients to Video: More consistency, creativity and control

2025 29

  1. Google's year in review: 8 areas with research breakthroughs in 2025
  2. Gemini 3 Flash: frontier intelligence built for speed
  3. Gemma Scope 2: helping the AI safety community deepen understanding of complex language model behavior
  4. Improved Gemini audio models for powerful voice experiences
  5. Deepening our partnership with the UK AI Security Institute
  6. Strengthening our partnership with the UK government to support prosperity and security in the AI era
  7. FACTS Benchmark Suite: Systematically evaluating the factuality of large language models
  8. Engineering more resilient crops for a warming climate
  9. AlphaFold: Five years of impact
  10. Revealing a key protein behind heart disease
  11. Breeding Healthier Honeybees With AlphaFold
  12. Google DeepMind supports U.S. Department of Energy on Genesis: a national mission to accelerate innovation and scientific discovery
  13. How we’re bringing AI image verification to the Gemini app
  14. Build with Nano Banana Pro, our Gemini 3 Pro Image model
  15. Introducing Nano Banana Pro
  16. Start building with Gemini 3
  17. We’re expanding our presence in Singapore to advance AI in the Asia-Pacific region
  18. A new era of intelligence with Gemini 3
  19. AlphaEarth
  20. AlphaGenome: AI for better understanding the genome
  21. Gemini Robotics On-Device brings AI to local robotic devices
  22. How we're supporting better tropical cyclone prediction with AI
  23. Advancing Gemini's security safeguards
  24. AlphaEvolve
  25. Music AI Sandbox, now with new features and broader access
  26. Taking a responsible path to AGI
  27. Building secure AGI: Evaluating emerging cyber security capabilities of advanced AI
  28. Introducing Gemini Robotics and Gemini Robotics-ER, AI models designed for robots to understand, act and react to the physical world.
  29. Updating the Frontier Safety Framework

2024 26

  1. FACTS Grounding: A new benchmark for evaluating the factuality of large language models
  2. Google DeepMind at NeurIPS 2024
  3. GenCast predicts weather and the risks of extreme conditions with state-of-the-art accuracy
  4. Genie 2: A large-scale foundation world model
  5. Pushing the frontiers of audio generation
  6. New generative AI tools open the doors of music creation
  7. Demis Hassabis & John Jumper awarded Nobel Prize in Chemistry
  8. How AlphaChip transformed computer chip design
  9. Empowering YouTube creators with generative AI
  10. Our latest advances in robot dexterity
  11. AlphaProteo generates novel proteins for biology and health research
  12. Mapping the misuse of generative AI
  13. Gemma Scope: helping the safety community shed light on the inner workings of language models
  14. AI achieves silver-medal standard solving International Mathematical Olympiad problems
  15. Google DeepMind at ICML 2024
  16. Generating audio for video
  17. Looking ahead to the AI Seoul Summit
  18. Introducing the Frontier Safety Framework
  19. Watermarking AI-generated text and video with SynthID
  20. Google DeepMind at ICLR 2024
  21. The ethics of advanced AI assistants
  22. TacticAI: an AI assistant for football tactics
  23. A generalist AI agent for 3D virtual environments
  24. AlphaGeometry: An Olympiad-level AI system for geometry
  25. Shaping the future of advanced robotics
  26. Images altered to trick machine vision can influence humans too

2023 15

  1. 2023: A Year of Groundbreaking Advances in AI and Computing
  2. FunSearch: Making new discoveries in mathematical sciences using Large Language Models
  3. Google DeepMind at NeurIPS 2023
  4. Millions of new materials discovered with deep learning
  5. Transforming the future of music creation
  6. Empowering the next generation for an AI-enabled world
  7. GraphCast: AI model for faster and more accurate global weather forecasting
  8. A glimpse of the next generation of AlphaFold
  9. Evaluating social and ethical risks from generative AI
  10. Scaling up learning across many different robot types
  11. A catalogue of genetic mutations to help pinpoint the cause of diseases
  12. Identifying AI-generated images with SynthID
  13. YouTube: Enhancing the user experience
  14. Google Cloud: Driving digital transformation
  15. MuZero, AlphaZero, and AlphaDev: Optimizing computer systems

2022 16

  1. Competitive programming with AlphaCode
  2. Stopping malaria in its tracks
  3. Fighting osteoporosis before it starts
  4. Understanding the faulty proteins linked to cancer and autism
  5. Solving the mystery of how an ancient bird went extinct
  6. Targeting early-onset Parkinson’s with AI
  7. Creating plastic-eating enzymes that could save us from pollution
  8. Advancing discovery of better drugs and medicine
  9. Accelerating the race against antibiotic resistance
  10. AlphaFold unlocks one of the greatest puzzles in biology
  11. AlphaFold transforms biology for millions around the world
  12. Tracing the evolution of proteins back to the origin of life
  13. The race to cure a billion people from a deadly parasitic disease
  14. Accelerating fusion science through learned plasma control
  15. MuZero’s first step from research into the real world
  16. DeepMind: The Podcast returns for Season 2

2021 3

  1. Simulating matter on the quantum scale with AI
  2. Language modelling at scale: Gopher, ethical considerations, and retrieval
  3. Exploring the beauty of pure mathematics in novel ways

2020 2

  1. Traffic prediction with advanced Graph Neural Networks
  2. Computational predictions of protein structures associated with COVID-19

2018 16

  1. Prefrontal cortex as a meta-reinforcement learning system
  2. Navigating with grid-like representations in artificial agents
  3. DeepMind, meet Android
  4. DeepMind papers at ICLR 2018
  5. Our first COO Lila Ibrahim takes DeepMind to the next level
  6. Retour à Paris / A return to Paris
  7. Learning to navigate in cities without a map
  8. Learning to write programs that generate images
  9. Understanding deep learning through neuron deletion
  10. Stop, look and listen to the people you want to help
  11. Learning by playing
  12. Researching patient deterioration with the US Department of Veterans Affairs
  13. Scalable agent architecture for distributed training
  14. Learning explanatory rules from noisy data
  15. Open-sourcing Psychlab
  16. Game-theory insights into asymmetric multi-agent games

2017 42

  1. 2017: DeepMind's year in review
  2. Collaborating with patients for better outcomes
  3. DeepMind papers at NIPS 2017
  4. Why doesn't Streams use AI?
  5. Specifying AI safety problems in simple environments
  6. Population based training of neural networks
  7. Applying machine learning to mammography screening for breast cancer
  8. High-fidelity speech synthesis with WaveNet
  9. Sharing our insights from designing with clinicians
  10. Bringing Streams to Yeovil District Hospital NHS Foundation Trust
  11. AlphaGo Zero: Starting from scratch
  12. Strengthening our commitment to Canadian research
  13. WaveNet launches in the Google Assistant
  14. Why we launched DeepMind Ethics & Society
  15. The hippocampus as a predictive map
  16. DeepMind and Blizzard open StarCraft II as an AI research environment
  17. DeepMind papers at ICML 2017 (part three)
  18. DeepMind papers at ICML 2017 (part two)
  19. DeepMind papers at ICML 2017 (part one)
  20. AI and Neuroscience: A virtuous circle
  21. Going beyond average for reinforcement learning
  22. Agents that imagine and plan
  23. Imagine this: Creating new visual concepts by recombining familiar ones
  24. Producing flexible behaviours in simulated environments
  25. DeepMind expands to Canada with new research office in Edmonton, Alberta
  26. Independent Reviewers release first annual report on DeepMind Health
  27. The Information Commissioner, the Royal Free, and what we’ve learned
  28. Interpreting Deep Neural Networks using Cognitive Psychology
  29. Enhancing patient safety at Taunton and Somerset NHS Foundation Trust
  30. Learning through human feedback
  31. A neural approach to relational reasoning
  32. AlphaGo's next move
  33. Exploring the mysteries of Go with AlphaGo and China's top players
  34. Innovations of AlphaGo
  35. Open sourcing Sonnet - a new library for constructing neural networks
  36. Distill: Communicating the science of machine learning
  37. Enabling Continual Learning in Neural Networks
  38. Trust, confidence and Verifiable Data Audit
  39. A milestone for DeepMind Health and Streams
  40. Understanding Agent Cooperation
  41. Our collaborations with academia to advance the field of AI
  42. DeepMind’s work in 2016: a round-up

2016 18

  1. Bringing the best of mobile technology to Imperial College Healthcare NHS Trust
  2. DeepMind Papers @ NIPS (Part 3)
  3. DeepMind Papers @ NIPS (Part 2)
  4. Open-sourcing DeepMind Lab
  5. DeepMind Papers @ NIPS (Part 1)
  6. Working with the NHS to build lifesaving technology
  7. Reinforcement learning with unsupervised auxiliary tasks
  8. DeepMind and Blizzard to release StarCraft II as an AI research environment
  9. Differentiable neural computers
  10. Announcing the Partnership on AI to Benefit People & Society
  11. Putting patients at the heart of DeepMind Health
  12. WaveNet: A generative model for raw audio
  13. Applying machine learning to radiotherapy planning for head & neck cancer
  14. Decoupled Neural Interfaces Using Synthetic Gradients
  15. DeepMind AI Reduces Google Data Centre Cooling Bill by 40%
  16. Deep Reinforcement Learning
  17. Announcing DeepMind Health research partnership with Moorfields Eye Hospital
  18. We are very excited to announce the launch of DeepMind Health

Undated 154

  1. Using ai to predict retinal disease progression
  2. Specification gaming the flip side of ai ingenuity
  3. Towards understanding glasses with graph neural networks
  4. Agent57 outperforming the human atari benchmark
  5. A new model and dataset for long range memory
  6. Dopamine and temporal difference learning a fruitful relationship between neuroscience and ai
  7. Alphafold using ai for scientific discovery 2020
  8. Using wavenet technology to reunite speech impaired users with their original voices
  9. Learning human objectives by evaluating hypothetical behaviours
  10. From unlikely start up to major scientific organisation entering our tenth year at deepmind
  11. Advanced machine learning helps play store users discover personalised apps
  12. Alphastar grandmaster level in starcraft ii using multi agent reinforcement learning
  13. Causal bayesian networks a flexible tool to enable fairer machine learning
  14. Deepminds health team joins google health
  15. The podcast episode 8 demis hassabis the interview
  16. The podcast episode 7 towards the future
  17. Replay in biological and artificial neural networks
  18. The podcast episode 6 ai for everyone
  19. The podcast episode 5 out of the lab
  20. The podcast episode 4 ai robot
  21. The podcast episode 3 life is like a game
  22. The podcast episode 2 go to zero
  23. The podcast episode 1 ai and neuroscience the virtuous circle
  24. Welcome to the deepmind podcast
  25. Using machine learning to accelerate ecological research
  26. Using ai to give doctors a 48 hour head start on life threatening illness
  27. How evolutionary selection can train more capable self driving cars
  28. Unsupervised learning the curious pupil
  29. Capture the flag the emergence of complex cooperative agents
  30. Identifying and eliminating bugs in learned predictive models
  31. Tf replicator distributed machine learning for researchers
  32. Machine learning can boost the value of wind energy
  33. Alphastar mastering the real time strategy game starcraft ii
  34. Alphazero shedding new light on chess shogi and go
  35. Alphafold using ai for scientific discovery
  36. Scaling streams with google
  37. Predicting eye disease with moorfields eye hospital
  38. Open sourcing trfl a library of reinforcement learning building blocks
  39. Expanding our research on breast cancer screening to japan
  40. Preserving outputs precisely while adaptively rescaling targets
  41. Using ai to plan head and neck cancer treatments
  42. Safety first ai for autonomous data centre cooling and industrial control
  43. A major milestone for the treatment of eye disease
  44. Objects that sound
  45. Measuring abstract reasoning in neural networks
  46. Deepmind papers at icml 2018
  47. Deepmind health response to independent reviewers report 2018
  48. Neural scene representation and rendering
  49. Royal free london publishes findings of legal audit in use of streams
  50. Unsupervised deep learning identifies semantic disentanglement in single inferotemporal face patch neurons
  51. On the expressivity of markov reward
  52. Improving language models by retrieving from trillions of tokens
  53. In conversation with ai building better language models
  54. From motor control to embodied intelligence
  55. Creating interactive agents with imitation learning
  56. Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents
  57. Red teaming language models with language models
  58. Probing image language transformers for verb understanding
  59. Learning robust real time cultural transmission without human data
  60. Gophercite teaching language models to support answers with verified quotes
  61. An empirical analysis of compute optimal large language model training
  62. Active offline policy selection
  63. A generalist agent
  64. Emergent bartering behaviour in multi agent reinforcement learning
  65. Dynamic language understanding adaptation to new knowledge in parametric and semi parametric models
  66. Evaluating multimodal interactive agents
  67. Perceiver ar general purpose long context autoregressive generation
  68. Unlocking high accuracy differentially private image classification through scale
  69. Byol explore exploration with bootstrapped prediction
  70. Intuitive physics learning in a deep learning model inspired by developmental psychology
  71. Human centred mechanism design with democratic ai
  72. Advancing conservation with ai based facial recognition of turtles
  73. Discovering when an agent is present in a system
  74. Alphafold reveals the structure of the protein universe
  75. Deepminds latest research at icml 2022
  76. Bridging deepmind research with alphabet products
  77. Kyrgyzstan to kings cross the star baker cooking up code
  78. Building a culture of pioneering responsibly
  79. Open sourcing mujoco
  80. From lego competitions to deepminds robotics lab
  81. When a passion for bass and brass help build better tools
  82. Tackling multiple tasks with a single visual language model
  83. Deepminds latest research at iclr 2022
  84. Real world challenges for agi
  85. Predicting the past with ithaca
  86. Opening up a physics simulator for robotics
  87. Stacking our way to more general robots
  88. Predicting gene expression with ai
  89. Nowcasting the next hour of rain
  90. Building architectures that can handle the worlds data
  91. Generally capable agents emerge from open ended play
  92. Putting the power of alphafold into the worlds hands
  93. Advancing sports analytics through ai research
  94. Game theory as an engine for large scale data analysis
  95. Muzero mastering go chess shogi and atari without rules
  96. Using jax to accelerate our research
  97. Alphafold a solution to a 50 year old grand challenge in biology
  98. Ferminet quantum physics and chemistry from first principles
  99. Fast reinforcement learning through the composition of behaviours
  100. Rt 2 new model translates vision and language into action
  101. Google deepmind research at icml 2023
  102. Using ai to fight climate change
  103. Codoc developing reliable ai tools for healthcare
  104. Exploring institutions for global ai governance
  105. Robocat a self improving robotic agent
  106. Alphadev discovers faster sorting algorithms
  107. An early warning system for novel ai risks
  108. Deepminds latest research at iclr 2023
  109. How can we build human values into ai
  110. Announcing google deepmind
  111. Ai for the board game diplomacy
  112. Mastering stratego the classic game of imperfect information
  113. Deepminds latest research at neurips 2022
  114. Building interactive agents in video game worlds
  115. Benchmarking the next generation of never ending learners
  116. Best practices for data enrichment
  117. How undesired goals can arise with correct rewards
  118. Measuring perception in ai models
  119. Discovering novel algorithms with alphatensor
  120. Making efficient use of demonstrations to solve hard exploration problems
  121. Restoring ancient text using deep learning a case study on greek epigraphy
  122. International evaluation of an ai system for breast cancer screening
  123. Artificial intelligence values and alignment
  124. Visual grounding in video for unsupervised word translation
  125. Learning to segment actions from observation and narration
  126. Simple sensor intentions for exploration
  127. Acme a new framework for distributed reinforcement learning
  128. Dm control software and tasks for continuous control
  129. Building safer dialogue agents
  130. How our principles helped define alphafolds release
  131. Rl unplugged benchmarks for offline reinforcement learning
  132. Using unity to help solve intelligence
  133. Imitating interactive intelligence
  134. Data architecture or losses what contributes most to multimodal transformer success
  135. Melting pot an evaluation suite for multi agent reinforcement learning
  136. Enabling high accuracy protein structure prediction at the proteome scale
  137. Challenges in detoxifying language models
  138. Maximising the impact of our breakthroughs
  139. Is curiosity all you need on the utility of emergent behaviours from curious exploration
  140. Bringing ai to the next generation of fusion energy
  141. Introducing codemender an ai agent for code security
  142. Gemini robotics 15 brings ai agents into the physical world
  143. Strengthening our frontier safety framework
  144. Discovering new solutions to century old problems in fluid dynamics
  145. Gemini achieves gold medal level at the international collegiate programming contest world finals
  146. Genie 3 a new frontier for world models
  147. Using ai to perceive the universe in greater depth
  148. How ai is helping advance the science of bioacoustics to save endangered species
  149. Aeneas transforms how historians connect the past
  150. Advanced version of gemini with deep think officially achieves gold medal standard at the international mathematical olympiad
  151. Exploring the context of online images with backstory
  152. Mapping modeling and understanding nature with ai
  153. Teaching ai to see the world more like we do
  154. Sima 2 an agent that plays reasons and learns with you in virtual 3d worlds