Lightnews — Scholar-powered news

Reposted by Ayush Bharti

ELLIS Institute Finland @ellisinstitute.fi · 2d

Thinking about a #PhD in AI + #machinelearning?

🇫🇮 Read about @ayushbharti.bsky.social's experience doing research in Finland:
www.ellisinstitute.fi/collaboratio...

Then,
👀 Apply for the @ellis.eu PhD program by October 31 🎃 : ellis.eu/news/ellis-p...

“Collaboration is a feature of the Finnish research community” | ELLIS Institute Finland

Strong collegial network is an asset to Ayush Bharti's research on robust machine learning.

www.ellisinstitute.fi

1 4

Reposted by Ayush Bharti

Henri Pesonen @hpesonen.bsky.social · Sep 1

I'm looking for a Doctoral Researcher (PhD student) to work with me on simulation-based inference at Data Science Research Centre, Tampere University Check the link for details and send an application before October 10th.

tuni.rekrytointi.com/paikat/?o=A_...

Doctoral Researcher (simulation-based inference) / Väitöskirjatutkija (simulaatio-pohjainen päättely)

tuni.rekrytointi.com

8 8

Reposted by Ayush Bharti

François-Xavier Briol @fxbriol.bsky.social · Aug 28

Just finished delivering a course on 'Robust and scalable simulation-based inference (SBI)' at Greek Stochastics. This covered an introduction to SBI, open challenges, and some recent contributions from my own group.

The slides are now available here: fxbriol.github.io/pdfs/slides-....

1 9 34

Ayush Bharti @ayushbharti.bsky.social · Jul 3

Great initiative!

François-Xavier Briol @fxbriol.bsky.social · Jul 3

Today is the day of the pre-ICML event at UCL! Come check out the exciting work from academics, industry researchers, postdocs and PhD students from around London: sites.google.com/view/pre-icm... @statisticsucl.bsky.social @ucl-csml-ellis.bsky.social

Pre ICML @ London 2025

The goal of this meetup is to bring together students, researchers, and engineers from the greater London area to discuss machine learning research presented at International Conference on Machine Lea...

sites.google.com

2

Ayush Bharti @ayushbharti.bsky.social · May 2

Thank you so much! Glad you liked it.

Ayush Bharti @ayushbharti.bsky.social · May 2

Doing so saves hours of computation time for the radio propagation model without any degradation in performance. (5/5)

1

Ayush Bharti @ayushbharti.bsky.social · May 2

Sampling from the cost-aware proposal is done via rejection sampling, and self-normalised importance weights are used to target the SBI posterior. (4/5)

1 3

Ayush Bharti @ayushbharti.bsky.social · May 2

We propose to sample from a cost-aware proposal to encourage sampling from the cheaper parameterisations of the model. (3/5)

1 3

Ayush Bharti @ayushbharti.bsky.social · May 2

Oftentimes, this computational cost varies with the parameter value, as is the case with this model from wireless communications field where the cost increases linearly. (2/5)

1 1

Ayush Bharti @ayushbharti.bsky.social · May 2

Thread below:

Popular SBI methods such as Approximate Bayesian computation (ABC), neural posterior estimation (NPE) and neural likelihood estimation (NLE) require running the simulator thousands of times, which can be a computational bottleneck. (1/5)

1 1

Ayush Bharti @ayushbharti.bsky.social · May 2

Poster

1 2

Ayush Bharti @ayushbharti.bsky.social · May 2

"Cost-aware simulation-based inference" is accepted at AISTATS 2025.

Check out our poster #205 on Sunday May 4th in Hall A-E if you are in Phuket. Finland's rising star @huangdaolang.bsky.social will be there to assist you :D

arxiv.org/abs/2410.07930

@fxbriol.bsky.social @samikaski.bsky.social

Cost-aware simulation-based inference

Simulation-based inference (SBI) is the preferred framework for estimating parameters of intractable models in science and engineering. A significant challenge in this context is the large computation...

arxiv.org

2 5 18

Ayush Bharti @ayushbharti.bsky.social · Apr 23

Looking forward to this!

Umberto Picchini @upicchini.bsky.social · Apr 23

Reminder that the next OWABI seminar www.warwick.ac.uk/owabi is scheduled on Thursday the 24th April at 11am UK time. Our next speaker is @ayushbharti.bsky.social (Aalto University), who will talk about "Cost-aware simulation-based inference".

Reposted by Ayush Bharti

Paul Chang @mummitrollet.bsky.social · Mar 12

Multi-Head Latent Attention vs Group Query Attention: We break down why MLA is a more expressive memory compression technique AND why naive implementations can backfire. Check it out!

datacrunch.io @datacrunch.io · Mar 12

⚡️Multi-Head Latent Attention is one of the key innovations that enabled @deepseek_ai's V3 and the subsequent R1 model.

⏭️ Join us as we continue our series into efficient AI inference, covering both theoretical insights and practical implementation:

🔗 datacrunch.io/blog/deepsee...

DeepSeek + SGLang: Multi-Head Latent Attention

Multi-Head Latent Attention (MLA) improves upon Group Query Attention (GQA), enabling long-context reasoning models and wider adoption across open-source LLMs.

datacrunch.io

10 25

Ayush Bharti @ayushbharti.bsky.social · Feb 14

Looking forward to speaking at the @approxbayesseminar.bsky.social!

François-Xavier Briol @fxbriol.bsky.social · Feb 14

My collaborator @ayushbharti.bsky.social will be presenting our recently accepted AISTATS paper on 'cost-aware simulation-based inference' at the next One World ABI Seminar on the 27th February.

Full details of the seminar series: warwick.ac.uk/fac/sci/stat...
arXiv paper: arxiv.org/abs/2410.07930

One World ABI Seminar

warwick.ac.uk

6

Reposted by Ayush Bharti

Approximate Bayes Seminar @approxbayesseminar.bsky.social · Feb 14

Our next talk will be on Thursday the 27th February at 11am UK time. Our next speaker is Ayush Bharti (Aalto University), who will talk about "Cost-aware simulation-based inference". To receive the link, sign up here: listserv.csv.warwick...

1 5 11

Ayush Bharti @ayushbharti.bsky.social · Dec 20

Congrats 🎉

3

Reposted by Ayush Bharti

Liza Semenova @liza-semenova.bsky.social · Dec 19

If you are interested in doing a #PhD with me at Imperial College London and qualify as a home student, please reach out (before end of 2024)! Potential topics: spatial statistics, applied deep generative models, probabilistic programming and more.

5 7

Ayush Bharti @ayushbharti.bsky.social · Dec 10

Congratulations Matias!

4

Reposted by Ayush Bharti

Daolang Huang @huangdaolang.bsky.social · Dec 5

Optimizing decision utility in Bayesian experimental design is key to improving downstream decision-making.

Excited to share our #NeurIPS2024 paper on Amortized Decision-Aware Bayesian Experimental Design: arxiv.org/abs/2411.02064

@lacerbi.bsky.social @samikaski.bsky.social

Details below.

1 12 41

Reposted by Ayush Bharti

Luigi Acerbi @lacerbi.bsky.social · Nov 27

@huangdaolang.bsky.social just joined here and you should follow him if you are interested in probabilistic machine learning, (Bayesian) exp. design and AI-assisted decision making. Not to mention that he has *several* NeurIPS papers already while in his 3rd PhD year...

1 8

Ayush Bharti @ayushbharti.bsky.social · Nov 21

Looks very interesting. This goes to the top of my reading list 😀

1 2

Ayush Bharti @ayushbharti.bsky.social · Nov 19

Hi, I'd like to join if that's ok.

1

Ayush Bharti @ayushbharti.bsky.social · Nov 19

✋

1

Ayush Bharti @ayushbharti.bsky.social · Nov 19

Sure. Will do.

1