Chris J. Maddison Profile picture
Asst. Prof. in Machine Learning @UofT. @CIFAR_News AI Chair @VectorInst and #LongCOVID patient.
Jul 24 14 tweets 4 min read
What if the next medical breakthrough is hidden in plain text? Causal estimates drives progress but data is limited & RCTs slow. Introducing NATURAL: a pipeline for causal estimation from text data in hours, not years.
Paper:
Site: tinyurl.com/ppr29
tinyurl.com/web98
A figure comparing the type of data used in RCTs vs. our pipeline, NATURAL. It also include a table that shows that in 4 real-world settings, NATURAL estimates came within 3 percentage points of a real-world RCT. NATURAL in a nutshell: take an expert-designed study protocol → filter for social media posts that conform to the protocol → LLMs extract distributions over key variables → adapt classical techniques (e.g., IPW) for treatment effect estimates. Our pipeline leverages LLMs to curate data that can be plugged into natural language conditioned estimators for average treatment effects.
Dec 21, 2023 6 tweets 2 min read
I decided to write out a part of the story of my illness for my #LongCovidMoonshot holiday letter, which I am sending tomorrow to @RepRaulGrijalva, @SenMarkKelly, and @SenatorSinema. 🌼 A sketch of poppy flowers. Dear Representative Raúl Grijalva,  I am an Arizona expat living in Canada. My zip code growing up in Tucson was 85716. I have been suffering from disabling Long Covid following my first COVID infection. This holiday season, I would like to tell you my story.  I moved to Canada at the start of the pandemic. My partner was wrapping up her studies at a conservation school in a different city. I missed her, so I went out to visit. Towards the end of the visit we went on a walk over the downs to a local windmill.   On this walk, we found dried poppy straw with seeds. We collected the seeds, and...
Mar 30, 2023 8 tweets 4 min read
#TeamClots preprinted an important milestone in the microclot theory of Long Covid etiology. 🧵on the implications, as I understood them. TLDR: it looks like microclot burden will struggle as a standalone biomarker for LC; we should look deeper!

researchsquare.com/article/rs-273… First, why is this an important milestone? This paper uses imaging flow cytometry to quantify the microclot burden in patients, which, AFAIK, is the first time that microclots have been measured in an objective and statistically robust way.
Mar 29, 2023 4 tweets 1 min read
To be fully transparent here, I consider potatoes to be the second-most inferior starch (yuca takes the crown), and I think this is an objective assessment. The criteria that I use is, how enjoyable is it to eat without being smothered in fat and salt?
Feb 21, 2023 13 tweets 2 min read
I've heard privately from a few non-patients, who interact regularly with Long Covid patients on Twitter, that we can be unwelcoming, and, at times, vile, in our advocacy. I want to lay out some scattered thoughts on how I have come to think about my own advocacy, such as it is. First, while I believe advocacy should never be vile, I acknowledge that there are different approaches, different strategies. This how is I think about it.