Very honoured to be included in this year’s
#Time100AI
list of influential people in AI🎈And more so the humbling messages from so many people ❤️. The work on directing AI to social purpose remains, and the ownership and inclusion of global communities🌍
🥳 Our work on machine learning for medium range weather predictions is this week’s cover of Science🎄This is a fun area to work in and lots more scope for us as a field to support a range of decision-making needs. 🌍
Trained on four decades of historical data,
#GraphCast
from
@GoogleDeepMind
is an
#AI
model that predicts global weather with greater speed and accuracy compared with traditional approaches.
Learn more in Science's last issue of 2023:
Exited to share our new paper: 'Monte Carlo Gradient Estimation in Machine Learning', with
@elaClaudia
@mfigurnov
@AndriyMnih
. It reviews of all the things we know about computing gradients of probabilistic functions. 🐾Thread👇🏾
Excited to be giving an introduction to Bayesian Machine Learning tomorrow
@MLSS_Tuebingen
🥳. Going attempt to cover: topics in Bayesian theory; tricks for manipulating probabilities; topics in ethics and social impact; raising interest in Bayesian thinking. 🎲
#virtual
#MLSS2020
Excited to share a new paper on Decolonisation and AI. With the amazing
@png_marie
@wsisaac
🤩 we explore why Decolonial theory matters for our field, and tactics to decolonise and reshape our field into a Decolonial AI. Feedback please 🙏🏾 Thread 👇🏾
We've worked for a few years on the problem of nowcasting, making short-term weather predictions ☔️. I think this is a powerful example of what is possible with generative models 🚀🤩. Excited to share our paper in
@Nature
today🎉🥳.
Really excited to share our latest paper in
@nature
today on machine learning for health data to make early predictions of acute kidney injury. Has been an amazing journey over the last 2 years and with an amazing set of people.
Beautiful overview of Bayesian Methods in ML by
@shakir_za
at
#MLSS2020
. Left me pondering about many things beyond Bayesian Inference. Thank you Shakir🙏
Quote of the day: “The cyclist, not the cycle, steers.“🚴♀️
🎤 P-I:
🎤 P-II:
First post for 2018! Machine Learning Trick of the Day (7): Density Ratio Trick.
Somehow let 2 years pass and a long list of tricks are waiting. In this post, on the importance of statistical comparisons, density ratios and classification.
The videos from the 3 lectures I gave at
#MLSS2018
are online🌱. If you watch any part of the 4.5 hours leave me some feedback: what's missing, should be left out, reordered, etc
Video 1:
Video 2:
Video 3:
The attendees at
#mlss2018
Madrid were one of the most engaged I've ever had: so many questions and thoughts😻. I gave 3 lectures on 'Planting the Seeds of Probabilistic Thinking'. 🌱🎲 See the slides here:
After a year of intense effort we are making our way to Nairobi for the 3rd Deep Learning Indaba
@DeepIndaba
😻 There's a community of people who have given every part of themselves to Raise Africa's voice in AI 👉🏽 🦁
#SautiYetu
#DLIndaba2019
#Masakhane
I was extremely excited to give a short talk on 'Queering Machine Learning'. Queering is an approach for a more critical ML that is available to us all. Thank you
@QueerinAI
workshop
@icmlconf
#ICML2020
🙏🏾🏳️🌈🥳✊🏾. Video, text, slides here 👇🏾
ICLR2019 will be in New Orleans💃🏾We just announced the call for papers with the submission deadline on 27 September. Happy to be one of this year's programme chairs with Alex Rush,
@svlevine
& Karen Livescu. Don't forget to read the length instructions.
Excellent essay by Mike Jordan - one I'm sure we'll re-read many times. Liked this: "It is Wiener’s intellectual agenda that has come to dominate in the current era, under the banner of McCarthy’s terminology"
First time I'll be visiting Madrid ✈️! Looking forward to speaking at the Machine Learning Summer where I'll give 3 lectures with the aim of 'planting the seeds of probabilistic thinking'. 🌱🎲
#mlss2018
The attendees at
#mlss2018
Madrid were one of the most engaged I've ever had: so many questions and thoughts😻. I gave 3 lectures on 'Planting the Seeds of Probabilistic Thinking'. 🌱🎲 See the slides here:
So cool!!
@timnitGebru
in
@thecontinent_
Africans of the Year 🌍🎉 What a nice illustration and the corresponding article is great. Read the full piece at
Sunday classic paper: Hamming (1986), You and Your Research. Enduringly popular, moving and thought-provoking.
I haven't shared it before so thought it would make a good classic reading for a reflective summer Sunday. 🎈
Proud that today we announce the 2019
#IndabaX
meetings. Between March and May, Machine Learning will grow across Africa in 26 countries (from 13 last year)! Please support the work of
@DeepIndaba
if you can. African ML is strong and thriving 🌍
#webuildtogether
#masakhane
We are proud to announce the 2019
#IndabaX
events will take place in 26 countries across our continent! 🌍 They build leadership and strength in Machine Learning across our African continent. Please support their efforts 🎈
#webuildtogether
So excited that
@siminyu_kat
is one of this years
@techreview
Innovators Under 35 🎉🥳🤩 Kathleen is amazing in so many ways. Grateful for her commitment to
@DeepIndaba
. This is a great recognition for all the big things she has only just begun to do.
Doing a review of model-based RL methods. Somehow, the PILCO paper
@icmlconf
by
@mpd37
is 10 years old 😳🤩. Lots of new methods, but this paper is still shaping our thinking about data-efficient policy search!
New post📬 Machine Learning Trick of the Day (8): Instrumental Thinking. On errors-in-variables, instrumental variables and reminding ourselves to question the assumptions in our models & practice. Been in my drafts since February 😬 so glad to finish this
📢Opportunity: Research Scientist Intern, Sociotechnical AI🌟 Please share in your networks 🙏 Joining our sociotechnical group developing research/artefacts/processes focussing on the interplay between society and AI. Details here:
Latest Bayesian analysis, Polson and Sokolov offer a Bayesian perspective on Deep Learning. I wrote my own view on this a while ago (which I think has more precision, but I’m obviously biased 🧐)
#SundayClassicPaper
📜: McDermott (1976) 'Artificial Intelligence Meets Natural Stupidity'. As we critique our own field, it is useful to see what recurs from the critique of the past. The critique on 'Wishful Mnemomics' seems still relevant.
Excited to be at
#ICML2023
in Hawaii 🌺 Excited to give a keynote on Tuesday entitled ‘Machine Learning with Social Purpose’ 🎈We’ll be covering everyone from weather forecasting, statistical testing, sociotechnical systems, and global community building.
New post: Decolonising Artificial Intelligence. Decolonisation emphasises self-ownership and self-confidence. This is a topic that I think will become more prominent; I wanted to learn more and explore different ways of thinking and of the paths forward 🧱
An honour to give a keynote
@icmlconf
this year 🙏🏾 Video is now online and I’m eager to probe further into this topic of Machine learning with social purpose. Talk in 3 parts: earth systems, Sociotechnical, & global AI. Would love your views 🎈
#ICML2023
Our work on AI for medium-range weather forecasts is published in
@ScienceMagazine
today. 🎉 A lot more for us to do as a community in this area; we hope these advances will support the vital weather-dependent decision-making that happens each day.
The view that these views were repulsive *in retrospect* is common, but a false argument. In their own time, eugenics, slavery, empire were resisted. Challenging such myths requires from us all a more demanding relationship with history; part of the work needed in this moment.
So many important scientists (Fisher, Newton) had views that are repulsive or strange in retrospect. It makes me wonder what my ancestors believed or did. I also wonder which of my views and actions will be judged as horrifying by subsequent generations
Finished my last recording- today for the
#NeurIPS2020
workshop on retrospectives and surveys🥳. Had the most fun with this recording since it was all about writing, writing style, and about writing and surveys in ML. Going to be a fun workshop🎉
The .
@googleresearch
PhD fellowship awards have been extended to include Africa. Please encourage good applicants to apply! Next deadline is 19 January 2019.
I always enjoy making my contribution to this course at UCL; gives me an opportunity to rediscover my field and to experiment with new ways of explaining. Here's my lecture:
Today we are excited to release video recordings of lectures from "Advanced Deep Learning and Reinforcement Learning", a course on deep RL taught at
@UCL
earlier this year by DeepMind researchers:
Enjoy!
My 2019
#WritingReflections
✍️🏾: I look back each year on the things I explored in my writing. I wrote fewer but bigger (better?) papers this year, wrote scripts for talks and spoke in more styles, did community organising in more ways, was taught by so many ♥️ thread🐾
The question of where we locate our meetings is one that I have been asked most often of late, and one on many people's minds. I wrote this as a short exploration of my thinking; and continue to explore other view on these questions.
#SundayClassicPaper
📜: On the foundations of statistics, Hacking (1964)🎲. This was later expanded into a very comprehensive book, and remains a key source for us to think about the definitions of probability. It's unfortunately only on JSTOR:
See this thread. Dave Silver gave us a great gift at
@DeepIndaba
. He spoke, for the first time 😻, about his 10 Principles of Reinforcement Learning. A fantastic view into Dave's thinking and a great source of topics for debate!
Congratulations to Marc
@mpd37
for receiving the
@imperialcollege
@ICComputing
President's award for being a global thought leader, excellence in research, and committment to diversity and the university's teaching mission. Well done 👏🏾🤩🎉🍾💐
2018 Writing Reflections ✍🏾: I thought to add more to the series on ML tricks, and did the first one in January to think about statistical comparisons, and density ratio, a popular tool in our field.
A huge congratulations to
@MannyKayy
who had a successful viva yesterday 💪🏾🎉♥️, with a great thesis on 'Efficient methods and architectures for deep neural network sequence models'. Always grateful to be an examiner of such great talent. Congrats Dr Kahembwe Mbabazi 🤩.
#SundayClassicPaper
📜: A Decision-based View of Causality. Lots of work going on in causality, so useful to reflect on this view defining causation in terms of unresponsiveness, and on Howard-form influence diagrams (aistats95). 🍂🎃
Sunday Classic Paper: Science and Statistics, EP Box (1976). The first from Box with the phrase "all models are wrong". On Cookbookery and Mathematistry; as debates on alchemy and rigour ensue, Box's advice continues to ground us.
We are actively seeking sponsors for the 2019 Deep Learning Indaba - this year in Nairobi 🇰🇪 If you, your organisation, or anyone you know can support the Indaba and make an investment in the next generation of African AI talent, please get in touch 🌍
And so
#ICLR2019
ends👋🏽 New Orleans and the conference were fantastic. This process showed how much effort our whole community puts in to make it a success🙏🏾. Thanks to this team for shaping the process.
@harvardnlp
@svlevine
Karen Livescu,
@_beenkim
Graham Taylor,
@aliceoh
🤩
An amazing group are building a community to interrogate and develop a wide understanding of how we think about our field of AI. And now there is a website 🕸️. Thanks
@MannyKayy
@timnitGebru
@sabelonow
@Abebab
and so many others.
2017 Writing Reflections: Wrote this piece to explore meta-cognition: ways of building systems that learn from uncertainty, and connections to awareness and consciousness. Also used this series to explore art as imagery for scientific communication.
Took this last week off to rest and disconnect🌸. Spent lots of my week reading these great books 🤓 Not sure how i'll face my ignored-inbox tomorrow though 🙈 Need more book recommendations for next time-off week. 🌿
Decisions released 🎉 Congratulations to accepted papers; to those who we could not accommodate, we wish you success in your ongoing research. See our blog for the first of our reflections. See you soon in Ethiopia.
#ICLR2020
#OurHatata
🇪🇹🌍
Please spread in your circles 🙏🏾 Great intern role for anyone interested in weather and machine learning and in particular on cyclones and to be part of our weather ML group. Plus Ferran is amazing to work with 🚀🌍🌀
Passionate in AI for Science, generative models and GNNs/Transformers? I'm looking for a PhD student with experience in these areas for an internship
@GoogleDeepMind
to revolutionize cyclone forecasting. Apply at by EOD tomorrow (Dec 15) and ping me!
#ICML2020
was so great, and the virtual format shows how many new ways of connection we have 🥳. So grateful to the authors, workshops, socials, mentorship, and the work of
@icmlconf
OC working tirelessly in the background. 😍🖤🎉 Thanks
@haldaume3
Aarti
#virtual
#NeuripsNext
Gave a
@TEDx
talk
@TEDxLSTM
last Nov🎙Called it the 'Machinery of Grace' and covered topics on responsible tech, global imbalances in AI, technological colonialism, grassroots, diversity and joy🎉.
Video🎬: Text📜:
Thank you
@ruha9
for making the trip to Nairobi and to the
@DeepIndaba
and for an amazing talk about race, technology, and inspiring us to think differently😻.
#SautiYetu
#DLIndaba2019
Obligatory pic of us following book cover pose. So excited to have met in person 🥰
We've been working quietly for a few months to get everything together and excited to see all the great submissions come in 🎉. Looking forward to starting the reviewing process and to eventually see
#ICLR2020
in Ethiopia and in Addis. 🌍
Thank you Machine Learning community for getting
#ICLR2020
off to such a great start 🎉. Congratulations to everyone who submitted and for pushing our science to new heights 📈.
Read the papers: 2594 now online
Next phase: Bidding, instructions on Friday.
Sunday classic paper: Ullman (1984) on Visual Routines. One of the foundations of thinking in compositional and relational learning, diectic representations, hierarchy. Long, but you’ll recognise many of the ideas 🕸.
Manifold and spectral learning is one of the areas that we will gain a great deal from by incorporating into our ML practice.
@pfau
has an amazingly deep knowledge of the area, and has done us a great service with this tutorial 🤩. Can't wait to watch the recording! 🎥
It was an honor to lecture at the Machine Learning Summer School in Buenos Aires (
@MLSS2018BsAs
) alongside an illustrious list of speakers last week. Wonderful city, wonderful students. The slides from my talk on manifold and spectral learning are here:
Kicking off a series of blog posts that explore different views and reactions on Decolonial AI: in this 1st post, through history and poetry and story, the relevance of coloniality and the landscape of power.
Decolonial Foresight (1): Put the Fire Out 🔥
What an amazing scene from the
#IndabaXRwanda
today. They did so much: had a 100 attendees, launched the women in ML and DS Kigali, presented posters, did some coding. They built together.
#masakhane
Congrats
@FutureDrLubalo
@bysd_a
and everyone else. 🤩
Yoshua Bengio, one of the fathers of deep learning, has been elected as a Fellow of the Royal Society. His work on neural networks and machine translation helped bring about the AI revolution crucial to self-driving cars and facial recognition.
#RSFellows
Sunday classic paper: Feynman (1966) ‘What is Science?’: ‘the belief in the ignorance of experts’, and his thoughts on guessing then testing, intelligence, rather strange views of women, importance of teachers. How would each of us answer this question?