Discover and read the best of Twitter Threads about #ML

Most recents (24)

1/ I looked at $RLAY with its data so far and all the technology going on. They are still my top pick in #AI/#ML in drug discovery.
2/ They use so many technologies to understand how proteins are encoded and move while they function. They use #ML to sort through DNA encoded library screens.
3/ They are taking on some of the most toxic targets in oncology and developing game changing drugs. They have 3 drugs for FGFR2, PI3Ka and SHP2. The PI3Ka space could be as high at 150,000 patients.
Read 4 tweets
1/ I have been spending a lot of time with the market selling off trying to share and teach about investing and how to build a portfolio to diversify across assets and risks.
2/ I think its good to take time when the market is down to focus on improving our game for the next time around. Since I am a biotech focused investor, I would like to share with you the trends within the biotech space I think could propel the next decade of innovation.
3/ If you know me already, these won't be very new to you. My first big concept is using tech in biotech with developing #AI, #ML and #Automation to drive down the costs and improve the success of clinical development.
Read 18 tweets
I went back and looked at last year. After the JPM conference in January, I probably used the word #bubble at least 200 times talking about the $XBI, #crypto and #EV. Today, I think biotech is at fire sales prices with an opportunity down 60%. The others still have a ways to go.
From the peak:
$XBI is down over 62%
$BTC is down over 57%
$TSLA is down almost 40%

I think biotech offers the best opportunity as #crypto has no real world applications. $TSLA is a very strong and profitable company, but still grossly overvalue by a good 20% or more.
Most biotechs are trading below cash levels. I haven't see valuations this cheap since the bottom of the 202 pandemic and the lows of the 2015 Hillary crash.
Read 5 tweets
Recently, a Princeton postdoc posted a thread about a paper he had published with his PI and group in PNAS, which raised serious methodological and ethical concerns. With many others, I tweeted my views of these problems, and did so with strong language.

In doing so, I contributed to a massive Twitter pile-on against this work, which this *junior* research could only have felt directed directly against himself. He has now deleted his Twitter account. I cannot believe that this is a coincidence.
I deeply regret my part in the pile-on. Even if criticism is not aimed at the researchers (and much of it was), massive numbers of often senior researchers directing harsh words at one's research, calling it unethical and comparing it to blatant racism, can only be a traumatic
Read 8 tweets
Over the past several months, I’ve been doing a deep-dive into transformer language models #NLProc and their applications in #psychology for the last part of my #PhD thesis. 👩🏻‍🎓 Here are a few resources that I’ve found invaluable and really launched me forward on this project 🧵:
🏃🏻‍♀️💨 If you already have some data and you want a jump start to classify it all using #transformers, check out @maria_antoniak’s #BERT for Humanists/Computational Social Scientists Talk: Colab notebook and other info:…
🤓 I’m an absolute nerd for stats and experiment design in psych, but doing #ML experiments is very different. A clear, detailed (and reasonable length) course to get up to speed on train-test splits, hyperparameter tuning, etc. and do the very best work:…
Read 7 tweets
Taking a look at the #AI/#ML use in biotech:

This is all about using tech to lower the costs and improve the rates of success in drug development. We still see 90% of drugs that enter the clinic fail to reach commercial. They can cost over $1 billion do develop.
1/ The first company is $EVO. They have a suite of technology they use to help other companies develop drugs. They use both Transcriptomics and Proteomics along with AI and ML. They work with many partners and get a small royalty on the sales of products they help develop.
2/ I don't own them as I want companies with big potential. Getting just a small royalty isn't going to get a huge growth potential unless they have thousands of partners. This makes them lower risk, but also lower reward. That is good if you like that kind of company.
Read 19 tweets
Looking at a little strategy:

Sharing some strategy while the sector is still down.
1/ I pick science themes like the use of #AI or #ML in drug discovery. Then I go out and find every company I can in that space. I make myself a list and start doing my research. I slowly whittle down the crowd to 5 names that stand out.
2/ Then I go further into the companies with listening to webcasts so I get to know if the management knows their stuff. I dig into the science and data. I even read the 10K or S-1 and understand the balance sheet. Then I look at the indications and the potential.
Read 17 tweets
#GTC22 is just around the corner, and it is a fully online and free event. There are numerous incredible and valuable sessions. If you are an academic, researcher, or an early-career ML/AI aspirant, the following sessions might be of particular interest to you. 🧵 👇 1/7 Image
5 Steps to Building a Career in AI: You’ll hear from AI experts as they give you insight into their journey and discuss the top five most practical steps you can take when beginning a career in artificial intelligence. 2/7
Fighting Diseases with High-resolution GPU-accelerated Molecular Dynamics Simulations. The rise of multi-GPU systems allows us to perform long timescale simulations either on supercomputers or in the cloud while enabling increased simulation accuracy. 3/7
Read 7 tweets
Taking a Look at $RLAY:

This is my updated look at Relay Therapeutics.
1/ Management:

I think Sanjiv Patel is a very good CEO. You won't hear a lot of promises or hype from him. He seems to be the quite type. He lets the data speak for itself. So far, I think he has done an outstanding job.
2/ Science:

Relay is using #AI and #ML to take on some of the most toxic targets in oncology. These are targets like FGFR and PI3Ka. These are targets that have been very promising in the past, but were just to toxic to go after.
Read 10 tweets
🚨Job talk thread🚨

Title: What Can *Conformal Inference* Offer to Statistics?


Main points:
(1) Conformal Inference can be made applicable in many #stats problems
(2) There are lots of misconceptions about Conformal Inference
(3) Try it!

Conformal Inference was designed for generating prediction intervals with guaranteed coverage in standard #ML problems.

Nevertheless, it can be modified to be applicable in

✔️Causal inference
✔️Survival analysis
✔️Election night model
✔️Outlier detection
✔️Risk calibration

Misconceptions about conformal inference:

❌ Conformal intervals only have marginal coverage and tend to be wide
✔️ Conformal intervals w/ proper conformity scores achieve conditional coverage & efficiency (short length) if the model is correctly specified

Read 6 tweets
Ranking my Companies by Management:

By science theme and management.
#AI/#ML in biotech

1. $RLAY
2. $RXRX
3. $SDGR
4. $EXAI (have not heard him present yet)

1. $BPMC
2. $MRTX
3. $ERAS
4. $RVMD
Read 7 tweets
Top Tech in Biotech names.

Looking at top 3 companies using #AI, #ML, #Automation, #DeepLearning and #CAD in biotech drug discovery.
This is one space where I have a very hard time picking a favorite out of my top 3 of $SDGR, $RXRX and $EXAI. Each of these companies is using technology in a very powerful way to drive drug discovery.
It can take many years and over $1 billion in cost to developing a drug as 90% of new science fails to reach commercial. That is due to most of biotech being trial and error. By using technology, we can lower the cost and improve the chances of success.
Read 9 tweets
The top science themes in the $XBI and the companies to follow.

#AI/#ML based Drug Discovery
Under #AI/#ML based Drug Discovery


All are using AI, ML, automation and/or deep learning to drive drug discovery to lower costs and improve success.
Under #ProteinDegraders


All are using targeted protein degradation to target old and new targets. Could be far safer then TKI and other approaches.
Read 9 tweets
The Biggest GDPR Fines of 2021…
In my mind, there's no doubt that as politicians in countries all over the world see:
1. The challenge of scaling #enterprise #dataGovernance and #dataCompliance to keep up with #bigData
2. Exponential growth of the use of person #PII data in #BI business decisions and #ml/#ai
3. The fragile #cloud systems built for scalability and elasticity, more than #compliance and #dataprivacy
4. The lucrative fine system based on a "per case" policy, where every single user or every single incident (multiple PII per user) = $$$
Read 8 tweets
Responsible Machine Learning book! 🦄📚

Not every day you get to see such a creative and artistic #DataScience book 🤯. The Hitchhiker’s Guide to Responsible Machine Learning is an educational comic in the area of Responsible Machine Learning with #R 👇🏼🧵

#ML #rstats Image
This beautiful book was created by Przemyslaw Biecek, Anna Kozak, and Aleksander Zawada.

The code in the book is with #RStats, code snippets are available on Rmarkdown as well (links below 👇🏼) ImageImageImage
The book goes over some of the fundamental concepts of responsible machine learning from the first step of acquiring the data and preparing it and through exploration, modeling, and communication steps. The book leverages art and interactivity, making the reading a fun experience ImageImage
Read 6 tweets
Are you looking to start with 𝐊𝐮𝐛𝐞𝐟𝐥𝐨𝐰 🌈? check this step by step installation guide for Windows by Ashish Patel 🧵👇🏼…

#MLOps #ML #Kubernetes Image
Kubeflow is an open-source #MLOps tool that provides toolkits for the deployment of machine learning workflows on #Kubernetes 🚀. Supports core data science tools such as Jupyter notebooks, training ML models such as #TensorFlow, #PyTorch, #XGBoost, setting pipelines, etc
The guide covers:
- #Docker installation
- Installation of Choco
- Installation and setting of Hyper-V
- Installation of #MiniKube and #Kubectl
Read 5 tweets
H2O new release! 🚀🚀🚀

This week, H2O had a major release of their ML open-source library for #R and #Python, introducing two new algorithms, improvements, and bug fixing. ❤️👇🏼 🧵

#MachineLearning #ML #DeepLearning #rstats #DataScience #DataScientists
New algorithm (1/2):
✨ Distributed Uplift Random Forest (Uplift DRF) - The Uplift DRF is a tree-based algorithm that uses a Random Forecast classifier to estimate a treatment's incremental impact. See demo on the notebook ⬇️…
#randomforest #ML #UpLift
New algorithm (2/2):
✨ Infogram & Admissible Machine Learning - is a new tool for machine learning interpretability. More details are available on the algorithm doc ⬇️…
#machinelearning #ML
Read 5 tweets
Did you know that An Introduction to Statistical Learning (ISLR) book has an online course? 🎥 🌈❤️
@edXOnline is offering an online course by @Stanford University, following the book curriculum: 🧵 👇🏼

#rstats #Statistics #ML #datascience
The course instructors are two of the book authors - Prof. Trevor Hastie and Prof. @robtibshirani. While the book is based on #R some awesome people translate it to #python, #julialang, and other #Rstats flavors (see links on the comments below 👇🏼).
The course covers the following topics (aligned with the book curriculum):
Read 9 tweets
This is what targeted racial harassment looks like: nitpicking things like “smiled” or “wore clothes like a celebrity, as a celebrity” or “bought cookware.” It is part of the “gold digger” narrative.
This is common in #healthcare#professionalism” where there is hazing & bullying by racist &/or classist made up non-rules lacking standards…yet people are called “breaking the rules.”

It’s exactly how bullies operate: double standards, mislabel/smear, exclude, marginalize.
In the U.K., “BAME” includes the formerly colonized, thus making race constructs or power different from U.S. as latter had race-based chattel slavery.

Thing is, those formerly colonized have many who have internalized self hate & racism.…
Read 76 tweets
#Highlights2021 for me: our #survey on efficient processing of #sparse and compressed tensors of #ML/#DNN models on #hardware accelerators published in @ProceedingsIEEE.
RT/sharing appreciated. 🧵
Context: Tensors of ML/DNN are compressed by leveraging #sparsity, #quantization, shape reduction. We summarize several such sources of sparsity & compression (§3). Sparsity is induced in structure while pruning & it is unstructured inherently for various applications or sources. Various sources induce stru...Common structures of sparsi...
Likewise, leveraging value similarity or approximate operations could yield irregularity in processing. Also, techniques for size-reduction make tensors asymmetric-shaped. Hence, special mechanisms can be required for efficient processing of sparse and irregular computations.
Read 12 tweets
Generative Models of Brain Dynamics – A #review

🎁 After a year of work, our #Xmas present is freshly out: 🎄

Get a bird's eye view with the synthesis of >200 refs at the intersection of #ML, #DynamicalSystems, and #Neuroscience!

A thread…🧵 (1/5)
To navigate through the broad landscape of neurodynamics modeling approaches, we map them based on the scale of organization/granularity and level of abstraction ~ conceptual scope (2/5)
We both covered classic cornerstones and recent SOTA methods, a methodological spectrum from naturalistic to abstract, from data- to hypothesis-driven generative models (3/5)
Read 5 tweets
Anybody who says “the #data don’t lie” either is ignorant or manipulative or both. The data are merely a tool that must be used responsibly & ethically to try to approximate “the truth” …some of which is unmeasurable (yet?). There are MANY #datascience methods & varying results
Everybody gets super excited about this new #AI #ML #machinelearning technique or that

You cannot build a RELIABLE house with low #quality bricks

First, look at the building blocks… meaning, how the #data fields are even defined & how the data are obtained

Who defined them?
I can’t tell you how glad I am that I have done coursework at both @MITSloan AND @StanfordGSB - Former immerses you in a ton of hands on analysis & options for analytic techniques useful in a #datascience job. Latter steps back to frame questions, assess missing data, biases.
Read 13 tweets
@sverhulst opens by explaining the #AI strategy and #data infrastructure. He notes how NYC is becoming a leader on AI strategy and provides an interesting precedent for global AI governance.
Stefaan introduces:
* @npparikh, NYC Mayor’s Office of the CTO
* @jbowlesnyc, Center for an Urban Future
* @CummingsRenee, Residence at the U of Virginia
* @lisard, Berkman Klein Center for Internet & Society at Harvard University
* @ruchowdh Twitter
Neal explains that the primary audience for this panel is inward focused, for those in gov and those who interact with it.

The major components are educational (informing what AI is), describing ecosystem in NYC, and describing findings and opportunities.
Read 40 tweets
Going to break-down how easy it is to use #autoML and more specifically JADBio AutoML. If you need an account to try it out, head over to and grab a free Basic plan. Ready? #data #ML (1/16)
STEP 1: You start by creating a Project on JADBio and generating all your study #data. That could either be data that has been processed and normalized by a #Bioinformatician, or public data available in the known data repositories (2/16)
If you’re using software for #molecular diagnosis like our Partner’s @QIAGEN OmicSoft Lands platform, your data is ready to be uploaded on JADBio. #ML 3/16
Read 17 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!