Po Liu
@poliu2s
Followers
177
Following
216
Media
13
Statuses
568
It's my #Twitterversary! I have been on Twitter for 13 years, since 23 Aug 2009 (via @twi_age).
0
0
0
It's my #Twitterversary! I have been on Twitter for 12 years, since 23 Aug 2009 (via @twi_age).
0
0
0
a must read for anyone who are into immunology. MHC-II neoantigens shape tumour immunity and response to immunotherapy | Nature
nature.com
Nature - In a mouse tumour model, immunotherapy-induced rejection of tumour cells requires presentation of both MHC class I and MHC class II antigens, which activate CD4+ and CD8+ T cells,...
0
52
120
Thank you @federicolois! For those interested, our CS234 reinforcement learning lecture notes and assignments are also available here:
The Reinforcement Learning course (CS234) by @emmabrunskill is pure gold. It's awesome that it is online.
4
55
236
It's my #Twitterversary! I have been on Twitter for 10 years, since 23 Aug 2009 (via @twi_age).
0
0
1
How to use the test data for model selection, from #icml2019. This could be very relevant for bioinformatics, where training and test datasets might differ substantially (eg. training on simulated data, testing on clinical/real data).
0
1
6
How do self-supervised features compare with standard features based on sequence alignment (PSSMs)? In general, self-supervised inputs produced fuzzier contact maps than PSSMs. This was a general theme: self-supervision shows promise but has room to improve. 5/8
2
2
5
Mutations in normal tissues - they are everywhere. But more so in the lung, esophagus and skin. In @sciencemagazine today https://t.co/8k6jhCueE5 Perspective by Tomasetti https://t.co/u4CFIAEYry
6
150
216
If you're wrapping up a PhD in computer science, statistics, math, comp biology, etc and looking to steer your energy to fight cancer, we're hiring! @DanaFarber @broadinstitute DM me or link:
1
49
61
The Convolutional Tsetlin Machine peaks at 99.51% accuracy on MNIST with a single layer of interpretable filters in…
1
31
121
A (brief) tweetorial (@CharlesSwanton request) for our latest work - Epigenetic evolution and lineage histories of chronic lymphocytic leukaemia
nature.com
Nature - A single-cell approach is used to follow the heritable stochastic changes to DNA methylation that occur in primary chronic lymphocytic leukaemia and healthy B cells, allowing the tracing...
3
43
118
2/2 Is data split for training fair? Where is comparison to state-of-art predictions for larger benchmark of mutation scan datasets? ( that are unsupervised from sequences alone) eg https://t.co/90OpdGg49c
1
7
15
My recent talk at @MIT which covers a lot of our exciting new work applying AI to chess, StarCraft 2, and protein folding is now available on youtube here: https://t.co/JCNRrlTRQU
#AlphaZero #AlphaGo #AlphaStar
12
316
1K
Machine learning predicts rapid relapse of triple negative breast cancer https://t.co/fG69cSfuaD
#biorxiv_cancer
0
2
4
Need to fine tune your #neuralnetwork to a specific task? Check out MorphNet, an open source technique for neural network model refinement that takes an existing network as input and produces a new one that is smaller and faster with better performance.
5
433
1K
The final version of my RGN paper is now online in @CellSystemsCP:
9
47
142
—OK, I'd like an atlas of breast #cancer at the single-cell level, including the immune system —No problem. Here it is. Nothing to it ;-) (just 194 patient samples, 26 million single cells) https://t.co/UYJBPmEHzK
@CellCellPress @CellPressNews by @BodenmillerLab
4
162
407
Today, we publicly released the data from Phase 1 of the MyHeart Counts study. See here for details.
nature.com
Scientific Data - Physical activity, sleep and cardiovascular health data for 50,000 individuals from the MyHeart Counts Study
1
20
47
The interpretation of genome data is quickly being transformed by #deeplearning #AI. And this is the best review on it yet, just out today https://t.co/ewrpHz4KgZ
@NatureRevGenet
@fabian_theis @gokcen @avsecz @gagneurlab
0
156
324