Andreas Madsen
@andreas_madsen
Followers
6K
Following
1K
Media
121
Statuses
2K
Ph.D. in NLP Interpretability from Mila. Previously, independent researcher + freelancer in ML and Node.js core developer.
Montreal, Canada
Joined December 2010
Spotlight award at ICML 2024. Introducing a new direction for interpretability: Faithfulness Measurable Models. 2x-5x better explanations, more accurate faithfulness metrics, no cost, and optimal explanations. Paper: Hall C 4-9, 24 Jul (11:30 - 13:00).
11
30
152
RT @tarrysingh: Check out the latest article in my newsletter: The Dragons Algorithm: I Saw the Future in China, and Its Already Happening….
linkedin.com
Hi everyone, I’m writing this having just returned from a whirlwind EU-China education trip, and to be frank, my perspective on the global technology landscape has been fundamentally altered. Our...
0
1
0
Python; calling it "t-string", is just troublesome when I need to lookup documentation at work.
0
0
3
0
7
0
RT @sarahookr: Very special being back in Montreal ✨. Huge thank you to @hugo_larochelle @apsarathchandar Marc-Alexandre Cote for bringing….
0
13
0
RT @kayo_yin: hearing “I’ve never really thought about signed language before but your talk made me want to work on it” is the biggest comp….
0
2
0
RT @apsarathchandar: In my lab, we have not one but four open postdoc positions! These positions cover developing foundation models for tex….
0
34
0
RT @1littlecoder: A new BERT baby! . If you are still using the huge RoBERTa or DeBERTa for your NLP tasks, here's NeoBERT! .
0
12
0
RT @Mila_Quebec: Congratuations to Lola (@lo_LB_La) and Sarath (@apsarathchandar) Read their blog post: https://t.c….
0
5
0
RT @apsarathchandar: What's the foundational model for generative chemistry? Our work, BindGPT, is a good candidate, and it will be present….
0
9
0
Also thanks to @apsarathchandar and @sivareddyg for supporting me during my Ph.D., which helped me get this far! I would highly recommend them if you are looking for a Ph.D. supervisor.
0
0
3
Positions:.* Full-stack.* Research Engineer.* Research Scientist.* Systems Infrastructure Engineer.* Research intern.Feel free to reach out but chances are I will see your application if you apply online. I will post details on my internship later, but there are more openings.
1
0
1
Excited to finally announce that I have joined @guidelabsai. We are building LLMs from scratch designed to be interpretable. Many have asked what I'm doing after my Ph.D., so great to finally get it out. We have a lot of open positions, from engineering or scientist to intern.
2
3
29
This is very sad news. @FelixHill84 reached out to me during my time as an independent researcher because he wanted to collaborate with me on some ideas. This was the first time anyone well-known had ever wanted to work with me, so I was deeply appreciative. Felix was always.
I’m really sad that my dear friend @FelixHill84 is no longer with us. He had many friends and colleagues all over the world - to try to ensure we reach them, his family have asked to share this webpage for the celebration of his life:
0
2
43
All investigations of faithfulness show that explanations' faithfulness is by default model and task-dependent. However, this is not the case when using FMMs. Thus, presenting a new paradigm for how to provide and ensure faithful explanations.
0
0
6
FMMs are when models are designed such that measuring faithfulness is cheap and precise, which makes it possible to optimize explanations toward maximum faithfulness.
1
0
8
Self-explanations are when LLMs explain themselves. Current models are not capable of this, but we suggest how that could be changed.
1
0
3
We ask the question: How to provide and ensure faithful explanations for general-purpose NLP models? The main thesis is that we should develop new paradigms in interpretability. The two new paradigms explored are faithfulness measurable models (FMMs) and self-explanations.
1
0
4
The full thesis is available at Thanks to @sivareddyg and @apsarathchandar for supervising me throughout all these years. It's been a great journey and I'm very grateful for their support.
arxiv.org
As machine learning becomes more widespread and is used in more critical applications, it's important to provide explanations for these models, to prevent unintended behavior. Unfortunately, many...
1
1
13
I’m thrilled to share that I’ve finished my Ph.D. at @Mila_Quebec. For the last 4.5 years, I have worked on creating new faithfulness-centric paradigms for NLP Interpretability. Read my vision for the future of interpretability in our newposition paper:
arxiv.org
Interpretability is the study of explaining models in understandable terms to humans. At present, interpretability is divided into two paradigms: the intrinsic paradigm, which believes that only...
25
42
400
I would highly recommend this lab! Consider joining if you are interested in NLP.
I have multiple vacancies for PhD and Masters students at @Mila_Quebec @McGill_NLP in NLP/ML focusing on representation learning, reasoning, multimodality and alignment. Deadline for applications is Dec 1st. More details:
0
0
7