We will present our poster titled "How Far Can Fairness Constraints Help Recover From Biased Data?"(with Amit Deshpande) @icmlconf on Wed 24 Jul 11:30 a.m., Hall C 4-9 #2513. Drop by our poster to discuss fair optimal classifiers and biased data:

RT @docmilanfar: A surprising & little-known results in classical statistics:. Mean (μ) and median (m) are within one std deviation:. |μ−m|….

RT @AnayMehrotra: We are organizing tutorial on Language Generation at #COLT 2025!. Visit our website (link below) for references and mater….

RT @LenaicChizat: Announcing : The 2nd International Summer School on Mathematical Aspects of Data Science.EPFL, Sept 1–5, 2025. Speakers:….

RT @behrouz_ali: ICML, I really prefer that someone ghost me, without acknowledging that they are ghosting me!.

I am hunting for an apartment in Tokyo for 8 months, starting April end. Want to keep it under 150,000 yen and at max 30 mins away from Nihombashi station. I would appreciate any references or help for this (and free pizza if you are in Delhi).

RT @xerefic: cupped some of this years harvest today . ft. @viggnesh90 @mohitsharmaj29 @mehulved

RT @AmartyaSanyal: Very shortly at @RealAAAI , @alexandrutifrea and I will be giving a Tutorial on the impact of Quality and availability o….

RT @simonguozirui: We asked Stanford students their best ML pickup lines 💌 for Valentine's Day 💝:. Check out the full dataset 🔍:. https://t….

RT @chatsagnik: 📜Our paper "Generalization Bounds for Dependent Data using Online-to-Batch Conversion" was accepted to AISTATS 2025! . In….

I am sure that @airindia will resolve this issue closer to my booking date and then claim that I will have to pay substantially extra since I am closer to the flying date. Thank you for the authentic @airindia experience, and I will make sure not to fly with you guys again. (3/3).

I have never had problems with Vistara customer care and dealing with @airindia only because of the merger. For once, I expected that @airindia would try to match Vistara's experience after the merger, but sadly, the story is even worse. (2/3).

Hey @airindia .I have been trying to reschedule my flight for over a month now without success. Due to the Vistara merger, I could not modify my booking through your website, and at this point, I have placed several callback requests without avail. (1/3).

We are very thankful to the reviewers, area chairs, and everyone involved in the review process @icmlconf. We had a very positive experience. Our paper is available on Arxiv here: Camera-ready and more details soon. (5/5).

Then, using threshold classifiers, we first reprove recovery results from Blum and Stangl (. Our technique can then generalize their results to arbitrary distributions and hypothesis classes (for example, using reject option classifiers). (4/5).

The advantage of such linear-fractional transforms is that under mild assumptions, they are order-preserving on the class probabilities and can be used to model complicated data bias scenarios like under-representation and label bias in many time steps. (3/5)

In this work, we expand upon the discussion on when fairness constraints can help recover the Bayes optimal classifiers. We first show that many popular data bias models can be expressed as a linear-fractional transform of the true class probability estimates. (2/5)

I am happy to announce that our paper titled "How Far Can Fairness Constraints Help Recover From Biased Data?" has been accepted at ICML 2024 (@icmlconf ). This is joint work with Amit Deshpande during my internship at Microsoft Research India (@MSFTResearch, @IndiaMSR). (1/5).

RT @chatsagnik: We are starting the 2024 edition of Talks on Quantum Computing at IIITD. This is going to be an online seminar series with….

RT @syrgkanis: Excited to share our new applied causal machine learning book is available online. Any feedback/cor….