Akshit @akshitwt

assessing ai capabilities. soon: ml grad @cambridge_uni. previously @precogatiiith, @iiit_hyderabad. futurebound. viciousaegis.github.io 22 Joined June 2023

Tweets

922
Followers

3K
Following

663
Likes

2K

Akshit @akshitwt

9 hours ago

me doing a masters to delay unemployment 🔥

JT @jiratickets

22 hours ago

me doing a masters to delay unemployment 🔥

172 86 5K 194K 281

1 0 13 721 0

Akshit @akshitwt

9 hours ago

my course hasnt even started yet and i already have my first practice sheet to finish wtf 😭

3 0 31 947 6

Download Image

fun fact, 2/3 of the members of iiit's ICPC team are actually precog members! @shiven_sinha who's been a veteran at CP for years and is a great researcher with a very cool paper lineup, and @HariAakashK, who is a *first-year* student(!) and was a co-first author in our recent…

Priyanshu Priyank @PriyanshuP1405

2 days ago

16 36 1K 51K 68

Download Image

6 3 113 4K 17

Akshit @akshitwt

3 days ago

i really don’t understand people who say papers should be written “like blogs” and should include stuff like real code there’s a reason mathematical equations and pseudocode exists. it’s makes everything formal and removes any ambiguity natural language may have. anyway,…

Daniel 🦔 @DanielW_Kiwi

3 days ago

960 266 10K 947K 2K

Download Image

3 2 55 2K 7

Chirag Agarwal @_cagarwal

3 days ago

We are excited to share some early progress in this direction for multimodal graph-language models (GLMs) and reasoning! Our analysis using interpretability tools shows that graph encoders are basically equivalent to a full GLM and expose modality bias in current GLM benchmarks.…

Akshit @akshitwt

a week ago

2 5 78 8K 24

Download Image

0 2 7 516 2

Akshit @akshitwt

3 days ago

my submission number for *a workshop* is in the 150s!! for context, any prior workshop i’ve submitted to has had max 40-50 submissions ml conferences are going through a crazy surge with the newfound hype in ai, will be interesting to see how they cope in the next few years

3 0 42 2K 1

Akshit @akshitwt

3 days ago

finalised a paper to submit on deadline day and found out openreview is down since morning 🤣

2 0 46 2K 1

Akshit @akshitwt

a week ago

LLMs can solve graph learning benchmarks... without access to a graph😲⁉️ Our new work highlights that explicitly injecting graph info into LLMs doesn't improve their performance on popular graph datasets. Is this a failure of benchmarking or the method? We show it's both 🧵👇