Sebastian

PhD Student at @uni_tue / IMPRS-IS (@MPI_IS). Prior ML/Data Science at @TU_Muenchen/@EPFL_en, @GeorgiaTech and @HPI_de

[Louis Kirsch @LouisKirschAI]

How can Transformers and other nets implement general-purpose learning algorithms in-context? Training on 2^14 tasks leads to strong generalization. Diff. to scaling laws, the memory size matters more than parameter count!

Meta-Learn Workshop #NeurIPS2022
https://t.co/P5a19Qodcz

[Intelligent Systems @MPI_IS]

Researching #AI #MachineLearning #Robotics #ComputerVision? Join our elite #PhD program - a partnership with @uni_stuttgart & @uni_tue! Applications accepted until Nov 1, 2022 at https://t.co/veVpmjl215 #Hiring #Job

[Jascha Sohl-Dickstein @jaschasd]

After 2 years of work by 442 contributors across 132 institutions, I am thrilled to announce that the https://t.co/wezEGzDEHt paper is now live: https://t.co/4Yg36EB9Ru. BIG-bench consists of 204 diverse tasks to measure and extrapolate the capabilities of large language models.

[Sebastian @fgewrgd]

@rasbt @roydanroy @tomgoldsteincs @shortstein You can do this with „Save page now“ at https://t.co/JMIXkFUbg1

[devistuia @devistuia]

Job opportunity! Come work with me and @b_kellenb at @EPFL_en in beautiful Sion and study how to map species and their interactions at scale with #DeepLearning !

Two phd positions available, info here https://t.co/CpqcHr6lP5

@snf_ch @epflENAC

[Sebastian @fgewrgd]

@epsiloncorrect @michael_galkin Are you sure you haven’t activated screen time synchronization with other Apple devices?

[Davide Mottin @DavideMottin]

Interested in #graphs and #ML? I have an open #postdoc position on a collaborative project with skilled industrial partners. Email me or apply at the link below.

https://t.co/fl3ksb4WlW

[A Schulz @SchulzAuguste]

I am super grateful and relieved that this is finally out - it's been a long process.🐌

In early 2019, we saw for the first time that our recurrent spiking neural network model could generate novelty responses solely through inhibitory plasticity.

A lot happened since then👇

[Richard Gao @_rdgao]

before Lisbon, I gave a @mackelab teatime talk about surfing as a beginner to prepare folks for the frustration

but it was actually about the PhD: in SD, I realized surfing as a beginner is a lot like starting a PhD, when you, invariably, also suck

this was my guide for both:

[Louis Kirsch @LouisKirschAI]

Excited to share our preprint 'Meta Learning Backpropagation And Improving It'

VS-ML meta learns general purpose learning algos that start to rival backprop in their performance. This is achieved purely by weight sharing LSTMs - no grads required. https://t.co/IKSG0TVAsO

👇1/7

[Sebastian @fgewrgd]

@_DaniBeck What is the memory pressure before that happens (Activity monitor)?

[Sebastian @fgewrgd]

@jeremychrysler @bradplumer @NeilShader The German postal service is using pretty much what you’re describing: https://t.co/XpA14gbFjn

[Sebastian @fgewrgd]

@sleepinyourhat Depending on your plans, you probably don’t even need an approval

[Sebastian @fgewrgd]

@sleepinyourhat Section 4 of https://t.co/7BOnIOmyoy is the interesting part. No commercial use and redistribution without authorization from the website.

[Sebastian @fgewrgd]

@sleepinyourhat Did you look into fanfiction (e.g. https://t.co/OIo76RTB46)? We built a dataset based on it in https://t.co/PQ0pvnwBDt

[Sebastian @fgewrgd]

@HayrullahDindar @oprahills @AcademicChatter @OpenAcademics Definitely try OCR and see if https://t.co/PO585tPq3L translates it better

[Sebastian @fgewrgd]

@victor_sergio Did you check Sebastian Stober‘s group in Magdeburg? https://t.co/Z2wiPNFR5t

[Jonathan Sauder 🇪🇺🇺🇦 @josauder]

New pre-print out! https://t.co/9cdY5kxO0v - we improve a state-of-the-art unrolled compressed sensing algorithm by employing a NN which adapts sparse reconstruction to the unknown target vector! 🔥 With @fidlojg and Peter Jung

[Davide Mottin @DavideMottin]

Stay tuned for tuned with the news of our new #NeurIPS2020 paper on how to *greatly* speed up complex matrix operations (e.g., matrix inversion) via a parallel, GPU-friendly, algorithm for SVD re-parameterisation!

[Sebastian @fgewrgd]

@jeremyphoward 8GB RAM for <10€: https://t.co/ohZ6wxTRO2