Nature Machine Intelligence

Nature Machine Intelligence has turned 5! Many thanks to all colleagues, authors and referees for helping us shape the journal. Read our anniversary edition of AI Reflections - interviews with recent Comment and Perspective authors

RT @NatureGeosci: Focus: Artificial intelligence in geoscience - collating recent @NatureGeosci research articles that use AI methods and opinion pieces on issues relating to the application of AI to geoscience

RT @ebervector: So pleased to have been able to write a little commentary piece with my advisor @weixx2 for @NatMachIntell! It's about this great work by @JamesGornet and Matt Thomson taking a look at how cognitive maps can arise just from predicting visual observations:

RT @SvenNyholm: New (short) paper by Silvia Milano (@SilviaMilano1) and me: "Advanced AI assistants that act on our behalf may not be ethically or legally feasible", in Nature Machine Intelligence (@NatMachIntell) #aiethics @MunichCenterML @LMU_Muenchen @aiatlmu

RT @QingxuZhu: I'm thrilled to announce that our paper has been published in Nature Machine Intelligence @NatMachIntell today! To encourage research on the natural and agile movement of quadrupedal robots, we've open-sourced our code and data. Check out our project page:

This year may see big advances in solving longstanding robotics challenges with generative AI - or are expectations too high? We discuss various viewpoints in our June editorial.

RT @hannahrosekirk: Great to see our personalised LLMs article in this @NatMachIntell editorial. Increased empathy exemplifies 2nd-order effects of personalised alignment...may seem preferable in the short-term but has long-term consequences for healthy human-AI interaction

RT @NatComputSci: 📢Our May issue is now live, and it includes a Perspective on computational frameworks for semiconductor discovery, a database for structure-based drug discovery, an algorithm to uncover laws of skill acquisition -- and much more! Check it out!

RT @pschwllr: More than a year after the preprint, I’m excited to have the first @SchwallerGroup study featured on the @EPFL landing page - out in @NatMachIntell! We present how LLM agents can be augmented with chemistry tools and demonstrate some of the few first successful syntheses — from user prompt to robotic platform execution. Thanks to the whole team! Let me highlight the two co-first authors, Andrés (@drecmb) and Sam (@SamCox822), and Andrew (@andrewwhite01), who were instrumental in this collaboration. Paper: Andrés’ thread: Andrew’s backstory (with pictures): @NCCR_Catalysis @EPFL_CHEM_Tweet @EPFL_AI_Center

EPFL researchers have developed ChemCrow, an AI system that enhances chemical research by integrating advanced tools for tasks like organic synthesis and drug discovery.

RT @Amit_Goldenb: Great questions about AI and empathy in this short piece: People cannot experience love from an LLM unless they act on the supposition that LLMs can love. Because LLMs cannot love, the experience of their love is premised on self-deception.

RT @jankosinski: This is the end of the world as we know it, if this is reproducible! The new era of functional modeling has begun. I took a transcription factor with an unknown structure and folded it with its recognition sequence embedded in longer DNA. AlphaFold3 accurately positioned the transcription factor.

RT @andrewwhite01: ChemCrow is out today in @NatMachIntell! ChemCrow is an agent that uses chem tools and a cloud-based robotic lab for open-ended chem tasks. It’s been a journey to get to publication and I’d like to share some history about it. It started back in 2022. 1/8

RT @BerruetaThomas: ⚡️New paper in @NatMachIntell⚡️ Embodied AI is the future, but are our algorithms ready for it? With MaxDiff RL, we reveal how continuity of experience breaks the performance of RL algorithms. Continuity is a fact of embodied experience, introducing correlations between data samples. MaxDiff RL was designed to account for this, leading to robust policies that work regardless of initial conditions or model variations---just look at the variance in our reward curves! MaxDiff RL also attains SOTA across embodied navigation benchmarks. In addition to our empirical results, we present a formal theory that we hope inspires future work examining the role of embodiment in AI. 👉Article: 👉Website:

RT @hannahrosekirk: Published in Nature Machine Intelligence today, our new article explores the trade-offs of personalised alignment in large language models ⚖️ Personalisation has potential to democratise decisions over how LLMs behave, but brings its own set of risks...

RT @igashov: Our method for molecular linker design with diffusion models (DiffLinker) got published in @NatMachIntell 🥳 Thanks a lot to our amazing coauthors @HannesStaerk Clément Vignac @rneschneuing @vgsatorras @pafrossard @wellingmax @mmbronstein @befcorreia

RT @adeenflinker: New paper out today in @NatMachIntell, where we show robust neural to speech decoding across 48 patients.

RT @skinniderlab: Very happy to share this work testing a widespread assumption in chemical AI and showing that invalid SMILES are a feature, not a bug:

RT @NatComputSci: 📢Our March issue is now live, and it’s a special one, including a Focus that highlights the state of the art, challenges, and opportunities in the development and use of digital twins across different domains. 👉 🧵1/11

Our March editorial, about the longstanding and evolving connections between neuroscience and AI research

RT @enzoferrante: We made it to the cover of Nature Machine Intelligence @NatMachIntell ! Congrats @rodbonazzola @affrangi and team! If you are interested in imaging genetics and how to discover new phenotype-genotype associations for anatomy, check it out: 👇

RT @ayushnoori: We’re pretty excited about this work from @KyleWSwanson, @james_y_zou, @ItsJonStokes and team, out now in @NatMachIntell. Kyle kindly shared the SyntheMol preprint and code with us a few months ago, and we’ve been able extend his work in some interesting new directions. 👀 1/2

I'm excited to share SyntheMol, a generative AI model for drug design optimized for creating easily synthesizable molecules. We applied our model to generate, synthesize, and validate 58 novel antibiotic candidates and found 6 potent hits. @NatMachIntell