Amir-massoud Farahmand@SoloGen

Vision: Understanding the computational and statistical principles required to design adaptive agents. Associate Prof @polymtl @Mila_Quebec 🇨🇦 #MahsaAmini

 Montreal, Quebec, Canada
academic.sologen.net/
Joined July 2008
134 Photos and videos
From a global equality and liberal perspective, it's quite great idea that everyone runs their own local LLM. This isn't possible with current AI architecture and hardware, but we can make it happen, as evident by the 20W devices in your home that are as smart as you are (*).
6
11
947
(*): Your brain! P.S: This is a reaction to certain CEO's utterances. If one day I become a CEO, I'd like to say things like above instead of "There is no reason for any individual to have a computer in his home" (Olsen) or "640K ought to be enough for anybody" (misatt to Gates).
1
164
I have never really got into Star Trek, but Captain Jean-Luc Picard has been an important part of the cultural fabric of our geekdom. Godspeed Patrick Stewart! 🖖
3
406
RT @tylerblack32: Disabled people should not live in poverty. It makes no sense the way we treat our people with disabilities in Canada. C…
82
13
Achievement unlocked: an Olympian (Donna Vakalis) cited my work! What a World Model Represents Is Three Questions (arxiv.org/abs/2607.06640) I haven't carefully read the paper to have a strong opinion, but it is interesting nonetheless!

What a World Model Represents Is Three Questions

World models learn task-relevant information through many routes: observation reconstruction, recurrent state, temporal filtering, and explicit task supervision. Different routes can make...

arxiv.org
3
3
35
2,761
I will also be recruiting students through ELLIS, either as the main advisor or co-advisor.
ELLIS@ELLISforEurope
Oct 1
🚨 Apply now for the #ELLISPhD Program! Work with leading AI/ML researchers across Europe, with exchange visits & joint supervision. The application deadline is 31 Oct. Get all info & apply now at ellis.eu/news/ellis-phd-prog…
1
12
1,207
Amir-massoud Farahmand retweeted
Glen Berseth@GlenBerseth
Oct 27
I'm hiring postdocs at Mila in Montreal: #reinforcementLearning, #robotics & #foundationmodels, and #automatedscience. Apply by Nov. 4.
1
5
42
3,299
Why do medical residents (and doctors, nurses, etc.) work very long shifts? Perhaps we should ask grad students to do the same, so they can keep up with the AI slop literature and keep their coding agents running contributing to the creation of new AI slops.
4
15
1,762
P.S.1: The suggestion is a joke. Don't fume! P.S.2: But I still don't understand the reason for residents: Don't we know that people's cognitive performance deteriorates after certain hours of work?
3
204
Amir-massoud Farahmand retweeted
Parisa Rashidi@Parisa__Rashidi
Sep 8
Can we improve LLM's clinical reasoning by teaching how clinicians reason instead of scaling them up? Excited to share our new arXiv preprint: 📄 Paper: arxiv.org/abs/2608.22622 💻 Code: github.com/iheallab/icureact 🤗 Dataset: huggingface.co/datasets/ihea… Models: huggingface.co/iheallab/Llam…
3
1
9
672
Now we will have a proliferation of tiger cubs 🐯 in future presentations! (The same way we did with 🍒 on top of 🧁!)
İsmail İlkan Ceylan@ismaililkanc
Oct 23
Hinton is giving the Gödel lecture at TU Wien and this is our current state according to him:
8
752
Do you have an AI-related idea for an start-up? Do you think if existing frontier lab's LLMs become 2x better, your product will be useless? Don't! P.S: I am not an entrepreneur; I cannot predict the future. Take my advice with a grain of salt.
3
395
Most (?) technicians don't know much about the techniques behind their tools! This was a great source of disappointment when I was a kid and I asked the electrician how the CRT in a TV worked, or why the plane flew (from a pilot), or how a nuclear reactor worked (from an uncle).
faith🌹@faith_vtch
Oct 19
I got my teeth 3d scanned at the dentist, I asked the lady if she could send me the 3D file. She looked at me like I had kicked a puppy and said “I’ve never heard of anyone asking for that…” like lady, it’s MY mouth. I want to 3d print my mouth as a gift to my friends. Not weird
1
5
633
I know this isn't the issue of the OP, but if they don't know, I cannot get the conversation to the point where they share things with me.
1
1
126
I haven't read it yet! I hope the "skepticism" in the title doesn't turn out to be "just because biological naturalism!". 🙄
Eric Schwitzgebel@eschwitz
Oct 18
My recent book AI and Consciousness: A Skeptical Overview is free online today and tomorrow (through Oct 19, presumably UK time). I don't think the hard copy is available yet. Get it while it's free! cambridge.org/core/elements/…
6
633
Do you want to be a part of the vibrant Mila community as an MS or a PhD student? Apply through Mila's supervision request process (deadline: Dec 1). We have so many wonderful PIs here working on diverse range of research topics. I will recruit 1-3 grad trainees this year.
Mila - Institut québécois d'IA
@Mila_Quebec
Oct 15
Mila's annual supervision request process is now open to receive MSc and PhD applications for Fall 2027 admission! For more information, visit mila.quebec/en/prospective-s…
1
17
1,566
I will be recruiting in areas close to RL Theory/Algorithm, on topics related to Model-based RL, RL Acceleration, Multi-agent RL, and Continual Learning. I have an extensive page for prospective students: academic.sologen.net/prospec…
1
3
5
394
If you think there is a good match between us, just add my name to your Mila application, and I will closely look at your profile.
2
134
Interested in RL, alignment between pre- and post-training, safe and verifiable agents, and learnability of data? Go work with Dylan!
This tweet is unavailable
1
26
1,693
Amir-massoud Farahmand retweeted
Majid Mohajerani@mohajerani77
Oct 14
1/ 🚀 We’re hiring a PhD student + a Postdoc to build foundation models for behaviour — where large-scale behavioural video meets deep learning. 📍 Douglas Research Centre / McGill + Mila, Montréal 🇨🇦 Apply 👉 l1nk.dev/j1zfg0k 🧵
1
4
13
1,378
Load more