Hoi To Wai

🌟Dive into the power of Stochastic Approximation (SA)! This mini-tutorial explores SA as a versatile tool for solving mean field equations—beyond gradient updates. 🔗Register at hubs.la/Q03zCH2-0. #MachineLearning #SignalProcessing #AIResearch #IEEE #AI

To be presented later today! @yauchungyiu

This afternoon!

RoSTE: An Efficient Quantization-Aware Supervised Fine-Tuning Approach for Large Language Models (icml.cc/virtual/2025/p…) with @quanwei0 & @yauchungyiu on Wed 16 Jul 11am-1:30pm East Hall (3/3).

Clipped SGD Algorithms for Performative Prediction: Tight Bounds for Clipping Bias and Remedies (icml.cc/virtual/2025/p…) with Qiang and @MichalYemini on Tue 15 Jul 4:30-7pm West Hall (2/n)

Next week at #ICML2025, we will be presenting two papers on #performativeprediction and #QAT training for LLM (1/n).

@afallah94 @RiceUniversity @RiceCompSci Congratulations!

I’m very excited to share that I’ll be joining @RiceUniversity as an Assistant Professor of Computer Science (@RiceCompSci) this fall! Huge thanks to my advisors and mentors throughout my PhD and postdoc!

Great work by my student Chenyue & Anthony's student Shangyuan! #graphsignalprocessing

Thank you all for attending today's tutorial on-site. Please feel free to reach out to us if you have any questions about Stochastic Approximation, and most importantly, enjoy ICASSP!

P.S. Due to visa issue, I will present the tutorial remotely. Gersende & Eric will be presenting from Hyderabad.

On Apr 7 AM (India time), we will be presenting a 3-hr tutorial @ieeeICASSP on "A Deep Dive into Recent Advances in Stochastic Approximation" with Gersende Fort & Eric Moulines. Check out the details here: sites.google.com/view/deep-dive… #stochasticapproximation

Great work by @yauchungyiu @quanwei0!

Our next talk will be given by Paolo Di Lorenzo on "Task-driven Topology Inference for Signal processing and Learning over Topological Domains". Join us on *Mar 5* (Wed) at *3pm CET*. Check out dsiseminar.github.io for details. #topologicallearning

@AryanMokhtari Congratulations Aryan!

Exciting News: I officially got tenure! Huge thanks to my amazing students, collaborators, and mentors!

Our next talk will be given by @ivodokmanic on "Understanding and improving graph neural nets with the cSBM". Join us on *Feb 26* (Wed) at *4pm CET* (New time!). Check out dsiseminar.github.io for details. #graphneuralnetworks

🚨 The 8th annual Graph Signal Processing Workshop is back this May 14-16! Held in Montreal, CA, at @Mila_Quebec, we’re covering all things graphs, signals, learning, & applications 🕸️〰️ 🔗: gspworkshop.org 👉🏻Abstract submission opens Feb 1 👉🏻 Registration opens Mar 20

@yenhuan_li What about stat.ML?

There are too many new papers in cs.LG on arXiv every day. I feel we need a new category, cs.LT, dedicated to learning theory. I personally do not consider learning theory to be a subfield of math.ST.

Second Unofficial CDC10k was a total success! @IEEECDC2024 @ceyhuneksin @hoitowai

My student, Chenyue, is presenting our LCSS paper on optimal prices (ieeexplore.ieee.org/abstract/docum…) in ~1hr at #CDCMilan2024 - new result showing the benefits of network-aware pricing strategy.

work done by my student Qiang Li and can be found here: openreview.net/forum?id=ejIzd…

check out our poster at #NeurIPS2024 this Thurs @ West Ballroom A-D #6205 on "Stochastic Optimization Schemes for Performative Prediction with Nonconvex Loss". TL;DR: with non-convex loss, SGD finds a biased stationary sol. & bias can be removed by lazy deployment.