rahulgk Profile Banner
Rahul G. Krishnan Profile
Rahul G. Krishnan

@rahulgk

Followers
2K
Following
2K
Statuses
701

Assistant Professor @UofTCompSci @LMP_UofT & @VectorInst ⚒️ 🏥 ML, AI, Causality and Medicine Also at https://t.co/8nRSO8DmJ7

Toronto,ON
Joined June 2009
Don't wanna be here? Send us removal request.
@rahulgk
Rahul G. Krishnan
2 months
I’ll be at @NeurIPSConf #NeurIPS2024 until Sunday. I'd love to chat about causality for medicine & science. I'm also looking for a postdoc interested in experimental design for medicine, if that's you, send me a message. I'll be presenting two papers at the main conference. 🧵(1/7)
1
4
48
@rahulgk
Rahul G. Krishnan
6 days
RT @narges_razavian: We built a 3D ViT foundation model on 361,663 clinical HeadCTs at NYU Medical center, and validated it on over 4 datas…
0
8
0
@rahulgk
Rahul G. Krishnan
8 days
RT @C_AnthonyMilton: I'm a Canadian journalist looking to interview Americans planning on moving to Canada in response to Trump's policies.…
0
6
0
@rahulgk
Rahul G. Krishnan
2 months
Time to start implementing better neural nets and algorithms to push the cost of combined pre+post training 🔽🔽🔽
@Sauers_
Sauers
2 months
The total compute cost was around $1,600,250, more than the entire prize
Tweet media one
0
0
9
@rahulgk
Rahul G. Krishnan
2 months
RT @VectorInst: Meet Aviraj Newatia, an undergrad student at @UofT and our Summer 2024 Research Intern! Working with Vector Faculty Member…
0
2
0
@rahulgk
Rahul G. Krishnan
2 months
RT @_allisonlau: Current LLM personalization methods can be costly and require multiple models. We introduce the Preference Pretrained Tran…
0
3
0
@rahulgk
Rahul G. Krishnan
2 months
RT @asaviaspossible: 1/6: Proud to present "Unlearning Tabular Data without a 'Forget Set'"! We introduce a new unlearning algorithm RELOA…
0
2
0
@rahulgk
Rahul G. Krishnan
2 months
Come by tomorrow to hear about some of the work we’re doing on generalizations of instruction tuning!
@YounwooC
Younwoo (Ethan) Choi
2 months
We explore how adding simple prompts during fine-tuning can improve LLM performance, offering a novel generalization of instruction tuning. Come to our poster #NeurIPS2024 Workshop on FITML 🗓️ SAT 14 Dec 4:40 - 5:30 PM PST 📍 East Exhibition Hall A
Tweet media one
0
0
5
@rahulgk
Rahul G. Krishnan
2 months
RT @rpatrik96: Today is the day! Come to East Hall in the afternoon to learn about how language models learn to compose language rules. @Ne
0
2
0
@rahulgk
Rahul G. Krishnan
2 months
RT @hamid_R_kamkar: Please attend our #NeurIPS2024 spotlight poster later today (11am-2pm) at the east exhibit hall #2809. Looking forwar…
0
5
0
@rahulgk
Rahul G. Krishnan
2 months
RT @vahidbalazadeh: How to use offline expert data with unobserved confounding to guide exploration in decision making? Our general approac…
0
4
0
@rahulgk
Rahul G. Krishnan
2 months
📷
0
0
2
@rahulgk
Rahul G. Krishnan
3 months
RT @colinraffel: Relu (yes, that ReLU), the superstar sysadmin who built the GPU servers behind AlexNet and many subsequent years of deep l…
0
25
0
@rahulgk
Rahul G. Krishnan
3 months
Someone should explain to him why this image is famous and what should really be done to get people moving.
Tweet media one
@fordnation
Doug Ford
3 months
When Toronto traffic is so bad that NHL teams have to walk to their games, we’ve got a serious problem. That’s why it’s so important that we deliver on our plan to fight gridlock by bringing sanity back to bike lane decisions, building highways and public transit, and speeding up construction on the Gardiner.
0
0
6