@kiaashour

DPhil student at @UniofOxford.

 Oxford, U.K.
kiaashour.github.io
Joined January 2023
  • Tweets321
  • Following518
  • Followers163
12Media
Pinned Tweet
Kia@kiaashourJul 9
1/n 🧵 Model evaluation is often bottlenecked by labelling cost. Active testing estimates a model's risk from far fewer labels, yet existing methods overlook a now-common resource: cheap predictions from powerful black-box models. We address this with PPAT.
3
7
1
25
3,967
No one changes the world alone Introducing Ampersand - the first model built for us. Ampersand is a collaborator trained to help people achieve their individual and collective goals. In multi-month simulations, users collaborating with Ampersand accomplished more than with Fable and other frontier models Available to everyone through our API
105
122
100
1,131
233,230
Kia retweeted
shafu
@shafu0xOct 8
crazy how a github profile doesn't mean anything anymore
112
116
27
4,268
137,761
Kia retweeted
arXiv.org@arxivOct 1
arXiv has updated our policy on rate limiting for all submitters. This update was made to fairly distribute moderator time & support the arXiv community of staff, volunteers, readers & authors. Please read our announcement to learn more: blog.arxiv.org/2026/10/01/up…
arXiv logo in white over a whiteboard with math symbols.

ALT arXiv logo in white over a whiteboard with math symbols.

79
496
239
1,730
760,165
RIP to the black and gold wandb
10
507
🔥Welcome 🔺Simplex Diffusion Models🔻! Standard discrete diffusion models suffer from "information collapse" because they discard uncertainty at intermediate steps by sampling categorical tokens. (1/5)
16
107
12
846
97,856
Students sometimes ask me if it still makes sense, in this accelerating age, to pursue a PhD in AI. Perhaps counterintuitively, I think it's a great time to do so. I wrote up some thoughts on this here: web.mit.edu/phillipi/www/wri…
92
584
109
3,863
1,742,752
Happy to share that PPAT has now been accepted to #NeurIPS2026! See you guys in Sydney/Paris.
Kia@kiaashourJul 9
1/n 🧵 Model evaluation is often bottlenecked by labelling cost. Active testing estimates a model's risk from far fewer labels, yet existing methods overlook a now-common resource: cheap predictions from powerful black-box models. We address this with PPAT.
1
4
30
1,011
yemen is in an ongoing famine right now and Facebook censored it, contributing to one of the worst humanitarian crises ever but interesting Alexandr chooses to joke about this in particular. human suffering probably not top of mind for Meta
Alexandr Wang
@alexandr_wangSep 24
“you must be a yemeni farmer addicted to khat if you think muse is going to have 1 billion users” muse:
137
66
19
1,401
151,205
Looks like we have over 500 submissions to ICLR this year. openreview.net/group?id=ICLR…

ICLR 2017

Welcome to the OpenReview homepage for ICLR 2017

openreview.net
13
108
50
654
interesting reversal happening where typos are now a sign of effort
106
372
65
4,388
135,698
ICLR limited first-time authors to just one submission this cycle. I've gotten multiple emails now to be added as an author on papers purely to bypass this limit. Senior authors were already added to papers purely for prestige; the rate of paper laundering is going to go up.
10
31
7
482
49,953
TMLR has faced a deluge of submissions, necessitating stricter desk rejection policies due to limited reviewer capacity Co-EiC Nihar Shah reached out to authors of 10 papers slated for desk reject. Could they answer questions about their *own* submission? medium.com/@TmlrOrg/asking-a…

Asking Authors About Their Own Papers

By Nihar B. Shah

medium.com
27
217
95
1,066
707,735
Thanks! 😀 Indeed, I solved neural self-reference (NS) in 1993 [1][2]. [1] JS. A self-referential weight matrix. ICANN 1993, Amsterdam, p 446-451 [2] K. Irie, I. Schlag, R. Csordas, JS. A Modern Self-Referential Weight Matrix That Learns to Modify Itself. ICML 2022
Florian Brand
@xeophonSep 8
there’s only one rightful author who can claim to have solved NS
22
52
18
1,145
114,404
#AISTATS 2027 Call for Paper is out! This year's edition is taking place in Montreal on May 3-6th, 2027 🇨🇦 Aymeric Dieuleveut and I are chairing virtual.aistats.org/Conferen… New this year: AI review, and more! Abstract deadline: September 29, 2026 Paper deadline: October 6, 2026
4
15
1
77
4,728
there’s only one rightful author who can claim to have solved NS
24
99
20
2,023
227,377
Auggie
@aug5thmusicSep 5
GPT-6 Astra has the best result yet on the Bach Benchmark. Its chorale contains no voice-leading errors, and its harmonic palette is sophisticated enough to include a Neapolitan sixth chord. More importantly, it is the first model to ever write passing tones on this benchmark, a significant leap in musical understanding. Extra High effort. Prompt below: “In LilyPond (version 2.24), write a 4-part chorale in the style of Bach, 3/4 time, G minor. Use two staves -- soprano and alto on the top staff, tenor and bass on the bottom staff. Respond with only the code block.”
72
787
46
12,200
914,117
Diffusion LLMs have two limitations relative to AR models: (1) Lower quality, and (2) Slower inference at large batch sizes. We address this "Uno" > Retains the AR architecture of LLMs > Each layer has two sets of weights: AR weights and Diffusion weights > Diffusion weights enable parallel sampling from the AR distribution losslessly Results: 💥Faster than all speculative decoding methods: DFlash and EAGLE-3 🔥 Beats ALL diffusion LLMs: Mercury 2, Diffusion Gemma, Llada Links to the paper, models, code below
30
82
15
388
65,896
“Why did ANOVA fall out of fashion?” statmodeling.stat.columbia.e…
3
40
1
230
39,014
Generative models can’t discover what they can’t reach. We’re excited to introduce ActFlow: a continued pre-training scheme that actively expands the valid design space reachable by flow and diffusion models. We call this generable set expansion — a new learning principle for out-of-distribution generative modeling, and a step toward evolvable search spaces for scientific discovery. (1/5)
12
58
5
454
43,016
New blog post: continuous diffusion for language is back! This research direction receded into the background for a while, but as of this year, it is once again a hot topic. I wrote down a historical perspective and some thoughts on the recent revival. sander.ai/2026/08/24/continu…

Continuous diffusion language models

Fully discrete methods dominated for a few years, but language models based on continuous diffusion are making a comeback.

sander.ai
22
139
20
779
96,703
Load more