Simon Vary

Fellow @MATSprogram

[Kia @kiaashour]

1/n
Want to calibrate your black-box predictor to have frequentist guarantees?
Have prior information you want to include in a robust, scalable way?
Then you might like our new paper: Robust Bayes-Assisted Conformal Prediction.

Joint work with Stefano Cortinovis & François Caron

[Iskander Azangulov @IAzangulov]

📄 Masked Language Flow Models (MLFMs)

We introduce a method to scale flow-based language models to reasoning and instruction-following tasks.

Joint work with @kiaashour @leoeleoleo1 @simonvary @prebeschini

🔗 https://t.co/DDnnfjwkcu

1/6

[Pierre Ablin @PierreAblin]

[Paper alert] "Optimization without Retraction on the Random Generalized Stiefel Manifold"

Introducing a cheap method to optimize functions on a *stochastic* manifold, defined by an expectation constraint.

https://t.co/WWtnWucjIa

See you in ICML :)

[Oxford Statistics @OxfordStats]

The Learning Theory & Statistics Optimization Reading Group meeting info:
📕Optimization with retraction on the random generalized Stiefel manifold.
🎙️Speaker: Šimon Váry
⌚Time: 4.15pm – 5.15pm
📍Location: Open Research Zone, Dept. of Stats
Open to #oxforduni UG, PGT, + PGR

[ISCS @ISCompSensing]

@SimonVary is giving an invited talk on exoplanet imaging at #ISCS23

[Pierre Ablin @PierreAblin]

Preprint alert!! https://t.co/urSGfZ5gek

"Infeasible Deterministic, Stochastic, and Variance-Reduction Algorithms for Optimization under Orthogonality Constraints"

w. P.-A. Absil, S.Vary, and B.Gao.

1/2

[Aleš Flídr @AlesFlidr]

Just found out that this is not common knowledge in CS: the term "bug" comes from the Lady beetleliteral insectsLady beetle messing up hardware. Here's a 1947 log of the Mark II computer.