Padhraic Smyth
@padhraicsmyth.bsky.social
ML researcher, teacher, professor, Dad, Irish, ...
Reposted by Padhraic Smyth
I was lucky to get a sneak preview of Advait’s talk and it is SO GOOD. I wish everyone building AI would watch this.
Microsoft senior researcher, Advait Sarkar, recently presented at TED AI Vienna, where he spoke about AI’s role in elevating human thought.
How to stop AI from killing your critical thinking
Chatbots might help you get work done faster — but at what cost? When we outsource our reasoning to artificial intelligence, we reduce ourselves to "middle managers for our own thoughts," says AI and ...
msft.it
November 13, 2025 at 5:57 PM
I was lucky to get a sneak preview of Advait’s talk and it is SO GOOD. I wish everyone building AI would watch this.
Reposted by Padhraic Smyth
The problem is us, with our Paleolithic vulnerabilities, our FOMO, our susceptibility to snake oil salesmen and the ELIZA effect. Say no to anthropomorphized tech solutionism and yes to stronger human institutions, fortified by ordinary technology. (11/11)
November 12, 2025 at 10:10 AM
The problem is us, with our Paleolithic vulnerabilities, our FOMO, our susceptibility to snake oil salesmen and the ELIZA effect. Say no to anthropomorphized tech solutionism and yes to stronger human institutions, fortified by ordinary technology. (11/11)
Reposted by Padhraic Smyth
“To leave our students to their own devices — which is to say, to the devices of AI companies — is to deprive them of...the means to understand the world they live in or navigate it effectively,” Anastasia Berg writes. www.nytimes.com/2025/10/29/o...
Opinion | Why Even Basic A.I. Use Is So Bad for Students
www.nytimes.com
October 31, 2025 at 6:15 PM
“To leave our students to their own devices — which is to say, to the devices of AI companies — is to deprive them of...the means to understand the world they live in or navigate it effectively,” Anastasia Berg writes. www.nytimes.com/2025/10/29/o...
Reposted by Padhraic Smyth
I cannot wait to celebrate TWENTY YEARS of #WiML at #NeurIPS in San Diego this December! 🎉🥳
Fun fact: The first #WiML was held in San Diego back in 2006! ❤️
Share your memories below, and come hang out on Dec 2! I will be there! I will be speaking! And have I mentioned I AM EXCITED?!?!?!
Fun fact: The first #WiML was held in San Diego back in 2006! ❤️
Share your memories below, and come hang out on Dec 2! I will be there! I will be speaking! And have I mentioned I AM EXCITED?!?!?!
WiML turns 20! 🎉
For two decades, WiML has connected, inspired, and supported women around the world who are advancing the field of ML. As we celebrate this incredible milestone, we want to hear your story. 💜
➡️ Share your reflections: shorturl.at/lxWM2
#WiML #WiML20Years #WiML20th
For two decades, WiML has connected, inspired, and supported women around the world who are advancing the field of ML. As we celebrate this incredible milestone, we want to hear your story. 💜
➡️ Share your reflections: shorturl.at/lxWM2
#WiML #WiML20Years #WiML20th
shorturl.at
October 31, 2025 at 9:09 PM
Reposted by Padhraic Smyth
Nice writeup in @caltech.edu news about the impact of the #Visipedia project in Computer Vision and Citizen Science
Pietro Perona's Vision: Visipedia and Its Lasting Impact on Computer Vision
The machine learning-driven system for identifying visual information has grown the citizen-science apps Merlin and iNaturalist, led to the development of key datasets, and jump-started the field of i...
www.eas.caltech.edu
October 24, 2025 at 9:48 PM
Nice writeup in @caltech.edu news about the impact of the #Visipedia project in Computer Vision and Citizen Science
Reposted by Padhraic Smyth
Not being an AI-doomer, but having experienced in my own department over the past 7 years a steady erosion of faculty governance norms and diminished prioritization of research-oriented pedagogy, this tracks. Let faculty and faculty interests lead the way, rather than administrators/regents
"AI returns to the university as part of a broad effort to further corporatize universities, vocationalize higher-education instruction, and diminish both research and research-based pedagogy."
October 24, 2025 at 3:33 PM
Not being an AI-doomer, but having experienced in my own department over the past 7 years a steady erosion of faculty governance norms and diminished prioritization of research-oriented pedagogy, this tracks. Let faculty and faculty interests lead the way, rather than administrators/regents
We (UC Irvine) are hiring for a faculty position (any level) in AI/ML/vision/NLP/etc. If you would like to work in a great department with great colleagues please apply! and please distribute to students, researchers, faculty who may be interested. More info here: drive.google.com/file/d/1ZyY5...
September 15, 2025 at 4:20 PM
We (UC Irvine) are hiring for a faculty position (any level) in AI/ML/vision/NLP/etc. If you would like to work in a great department with great colleagues please apply! and please distribute to students, researchers, faculty who may be interested. More info here: drive.google.com/file/d/1ZyY5...
Reposted by Padhraic Smyth
Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: web.stanford.edu/~jurafsky/sl...
Speech and Language Processing
Speech and Language Processing
web.stanford.edu
August 24, 2025 at 7:28 PM
Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: web.stanford.edu/~jurafsky/sl...
Reposted by Padhraic Smyth
📣 Please share: We invite submissions to the 29th International Conference on Artificial Intelligence and Statistics (#AISTATS 2026) and welcome paper submissions at the intersection of AI, machine learning, statistics, and related areas. [1/3]
August 12, 2025 at 11:46 AM
📣 Please share: We invite submissions to the 29th International Conference on Artificial Intelligence and Statistics (#AISTATS 2026) and welcome paper submissions at the intersection of AI, machine learning, statistics, and related areas. [1/3]
Reposted by Padhraic Smyth
Every rich person is going to tell *you* how great AI teaching is while sending *their* kids to the kind of schooling the Ancient Greeks would recognize. I just wish everyone would think about why that is.
I think we’re on the cusp of mass attempts to automate teaching, and I also think it is going to be a massive wasteful failure in ways that will make the “learning loss” of the pandemic look like a speed bump next to a mountain.
August 5, 2025 at 4:46 PM
Every rich person is going to tell *you* how great AI teaching is while sending *their* kids to the kind of schooling the Ancient Greeks would recognize. I just wish everyone would think about why that is.
Reposted by Padhraic Smyth
Our computer vision textbook is now available for free online here:
visionbook.mit.edu
We are working on adding some interactive components like search and (beta) integration with LLMs.
Hope this is useful and feel free to submit Github issues to help us improve the text!
visionbook.mit.edu
We are working on adding some interactive components like search and (beta) integration with LLMs.
Hope this is useful and feel free to submit Github issues to help us improve the text!
Foundations of Computer Vision
The print version was published by
visionbook.mit.edu
June 15, 2025 at 3:45 PM
Our computer vision textbook is now available for free online here:
visionbook.mit.edu
We are working on adding some interactive components like search and (beta) integration with LLMs.
Hope this is useful and feel free to submit Github issues to help us improve the text!
visionbook.mit.edu
We are working on adding some interactive components like search and (beta) integration with LLMs.
Hope this is useful and feel free to submit Github issues to help us improve the text!
Reposted by Padhraic Smyth
I've heard this personally from multiple PMs at AI companies. Students are one of the biggest demographics and they need to "break in" and have even more usage to improve their metrics. Classic corporate economic incentives
AI companies in the US gave access to their systems to students for free during college exams
China disabled access to AI systems during nationwide college exams www.theverge.com/news/682737/...
Feel free to draw your own conclusions
China disabled access to AI systems during nationwide college exams www.theverge.com/news/682737/...
Feel free to draw your own conclusions
China shuts down AI tools during nationwide college exams
New age problems require new age solutions.
www.theverge.com
June 9, 2025 at 3:24 PM
I've heard this personally from multiple PMs at AI companies. Students are one of the biggest demographics and they need to "break in" and have even more usage to improve their metrics. Classic corporate economic incentives
Reposted by Padhraic Smyth
This Senate proposal advanced by Sen Cruz would cancel new and existing State laws on any aspect of tech use including civil rights, consumer protection, privacy, fraud, safety for kids, accessibility, and more. In short, we’d lose the few laws we have that ensure responsible AI use. #killthebill
Don’t be confused: The Senate Republican AI language includes the EXACT SAME AI moratorium as the House bill. Almost word-for-word. This is a complete, 10-year ban on state AI regulation. Period.
June 6, 2025 at 4:29 AM
This Senate proposal advanced by Sen Cruz would cancel new and existing State laws on any aspect of tech use including civil rights, consumer protection, privacy, fraud, safety for kids, accessibility, and more. In short, we’d lose the few laws we have that ensure responsible AI use. #killthebill
Reposted by Padhraic Smyth
People keep plugging AI "Co-Scientists," so what happens when you ask them to do an important task like finding errors in papers?
We built SPOT, a dataset of STEM manuscripts across 10 fields annotated with real errors to find out.
(tl;dr not even close to usable) #NLProc
arxiv.org/abs/2505.11855
We built SPOT, a dataset of STEM manuscripts across 10 fields annotated with real errors to find out.
(tl;dr not even close to usable) #NLProc
arxiv.org/abs/2505.11855
May 23, 2025 at 4:22 PM
People keep plugging AI "Co-Scientists," so what happens when you ask them to do an important task like finding errors in papers?
We built SPOT, a dataset of STEM manuscripts across 10 fields annotated with real errors to find out.
(tl;dr not even close to usable) #NLProc
arxiv.org/abs/2505.11855
We built SPOT, a dataset of STEM manuscripts across 10 fields annotated with real errors to find out.
(tl;dr not even close to usable) #NLProc
arxiv.org/abs/2505.11855
Reposted by Padhraic Smyth
Llama 3.1 70B contains copies of nearly the entirety of some books. Harry Potter is just one of them. I don’t know if this means it’s an infringing copy. But the first question to answer is if it’s a copy at all/in the first place. That’s what our new results suggest:
arxiv.org/abs/2505.12546
arxiv.org/abs/2505.12546
Extracting memorized pieces of (copyrighted) books from open-weight language models
Plaintiffs and defendants in copyright lawsuits over generative AI often make sweeping, opposing claims about the extent to which large language models (LLMs) have memorized plaintiffs' protected expr...
arxiv.org
May 21, 2025 at 11:20 AM
Llama 3.1 70B contains copies of nearly the entirety of some books. Harry Potter is just one of them. I don’t know if this means it’s an infringing copy. But the first question to answer is if it’s a copy at all/in the first place. That’s what our new results suggest:
arxiv.org/abs/2505.12546
arxiv.org/abs/2505.12546
Reposted by Padhraic Smyth
A call for scientists to stand up for scientific freedom as well as funding: www.nature.com/articles/d41...
May 14, 2025 at 1:36 AM
A call for scientists to stand up for scientific freedom as well as funding: www.nature.com/articles/d41...
Reposted by Padhraic Smyth
@candicemorey.bsky.social and I were just talking about this. Students, I think, are still (rightly) nervous about submitting LLM-produced work, but they are using it to summarise papers they struggle to read. And it shows in their subsequent writing. It's "just reading the abstracts", but worse.
This paper makes important point that we're not talking about impact of AI on reading enough. Marking work that's full of AI is demoralising but in some ways, the threat AI poses to comprehension is bigger than the threat to what students produce...
www.tandfonline.com/doi/full/10....
#AcademicSky
www.tandfonline.com/doi/full/10....
#AcademicSky
May 13, 2025 at 7:16 AM
@candicemorey.bsky.social and I were just talking about this. Students, I think, are still (rightly) nervous about submitting LLM-produced work, but they are using it to summarise papers they struggle to read. And it shows in their subsequent writing. It's "just reading the abstracts", but worse.
Reposted by Padhraic Smyth
Because we must build good things while we scream about the bad, I have started a "Data for Good" team @data-for-good-team.bsky.social that partners with organizations needing short-term data science help. We have three projects ongoing & will add more as our capacity grows.
data-for-good-team.org
data-for-good-team.org
May 10, 2025 at 3:33 PM
Because we must build good things while we scream about the bad, I have started a "Data for Good" team @data-for-good-team.bsky.social that partners with organizations needing short-term data science help. We have three projects ongoing & will add more as our capacity grows.
data-for-good-team.org
data-for-good-team.org
Reposted by Padhraic Smyth
Congratulations to Rich Sutton and Andrew Barto on receiving the Turing Award in recognition of their significant contributions to ML. I also stand with them: Releasing models to the public without the right technical and societal safeguards is irresponsible.
www.ft.com/content/d8f8...
www.ft.com/content/d8f8...
Turing Award winners warn over unsafe deployment of AI models
Two pioneers of reinforcement learning have won the $1mn prize from the Association for Computing Machinery
www.ft.com
March 5, 2025 at 1:43 PM
Congratulations to Rich Sutton and Andrew Barto on receiving the Turing Award in recognition of their significant contributions to ML. I also stand with them: Releasing models to the public without the right technical and societal safeguards is irresponsible.
www.ft.com/content/d8f8...
www.ft.com/content/d8f8...
Reposted by Padhraic Smyth
My new paper "Deep Learning is Not So Mysterious or Different": arxiv.org/abs/2503.02113. Generalization behaviours in deep learning can be intuitively understood through a notion of soft inductive biases, and formally characterized with countable hypothesis bounds! 1/12
March 5, 2025 at 3:38 PM
My new paper "Deep Learning is Not So Mysterious or Different": arxiv.org/abs/2503.02113. Generalization behaviours in deep learning can be intuitively understood through a notion of soft inductive biases, and formally characterized with countable hypothesis bounds! 1/12
Reposted by Padhraic Smyth
There are a bunch of published papers referencing “vegetative electron microscopy” because of Generative AI retractionwatch.com/2025/02/10/v...
February 22, 2025 at 1:47 AM
There are a bunch of published papers referencing “vegetative electron microscopy” because of Generative AI retractionwatch.com/2025/02/10/v...
Reposted by Padhraic Smyth
I'm very excited to share notes on Probabilistic AI that I have been writing with @arkrause.bsky.social 🥳
arxiv.org/pdf/2502.05244
These notes aim to give a graduate-level introduction to probabilistic ML + sequential decision-making.
I'm super glad to be able to share them with all of you now!
arxiv.org/pdf/2502.05244
These notes aim to give a graduate-level introduction to probabilistic ML + sequential decision-making.
I'm super glad to be able to share them with all of you now!
February 11, 2025 at 8:19 AM
I'm very excited to share notes on Probabilistic AI that I have been writing with @arkrause.bsky.social 🥳
arxiv.org/pdf/2502.05244
These notes aim to give a graduate-level introduction to probabilistic ML + sequential decision-making.
I'm super glad to be able to share them with all of you now!
arxiv.org/pdf/2502.05244
These notes aim to give a graduate-level introduction to probabilistic ML + sequential decision-making.
I'm super glad to be able to share them with all of you now!
Reposted by Padhraic Smyth
What should we teach our undergrads about machine learning? I wrote up some ideas for restructuring Machine Learning 101.
Machine Learning 101?
Imagining a new syllabus for a first course on machine learning.
www.argmin.net
February 13, 2025 at 3:39 PM
What should we teach our undergrads about machine learning? I wrote up some ideas for restructuring Machine Learning 101.
Reposted by Padhraic Smyth
1/8. This week I had a few different conversations with scholars who, in the face of the attacks on science and institutions of learning in the U.S., are wondering what to do. One suggestion I have is: keep doing your work. It matters in and of itself. Why do I say that? A few reasons.
February 8, 2025 at 8:42 PM
1/8. This week I had a few different conversations with scholars who, in the face of the attacks on science and institutions of learning in the U.S., are wondering what to do. One suggestion I have is: keep doing your work. It matters in and of itself. Why do I say that? A few reasons.
Reposted by Padhraic Smyth
Microsoft's own research confirms something that was already pretty obvious: relying on a text generating machine to come up with answers erodes critical thinking, and is a method favoured by those who never liked doing critical thinking in the first place
advait.org/files/lee_20...
advait.org/files/lee_20...
advait.org
February 9, 2025 at 10:15 AM
Microsoft's own research confirms something that was already pretty obvious: relying on a text generating machine to come up with answers erodes critical thinking, and is a method favoured by those who never liked doing critical thinking in the first place
advait.org/files/lee_20...
advait.org/files/lee_20...