Timo Witten

Postdoc, math, AI, neuroscience. 🏳️‍🌈🇫🇷🇩🇪🇺🇸

[Timo Witten @TimoWitten]

Most talkative people on Twitter are actually no longer doing great science.

[Timo Witten @TimoWitten]

@behrenstimb Visit me, Tim. Hope that will help my manuscript get reviewed at eLife. ;)

[David Pfau @pfau]

@emukamel @KordingLab Neuro for the sake of understanding how the brain works is a perfectly lofty goal in its own right. It's just objectively true that current neuro research is doing little to nothing to drive AI research forward. Nothing wrong with that, they're just different fields.

[Luca Ambrogioni @LucaAmb]

@ylecun @pfau @KordingLab You you put an idea into math, it is easy to think that it is just math

The problem is that there are infinitely many ways to get something that us just math, not much of a guidance

[Timo Witten @TimoWitten]

Neuroscientists want to get money. In fact, they know very little about the brain to be able to inspire modern AI.

[Sam Gershman @gershbrain]

If you're wondering whether neuroscience is useful for AI, consider the counterfactual world where engineers knew nothing about neuroscience. Do you think that we wouldn't have convolutional networks or reinforcement learning?

[Timo Witten @TimoWitten]

@jakhmack @BrentDoiron “Trivial results”.

[Gary Marcus @GaryMarcus]

fantastic essay by @R3_GSI

[Tony Zador @TonyZador]

"Progress in science depends on new techniques, new discoveries and new ideas, probably in that order"
--- Sydney Brenner (~1980)

[Dr. Parik Patel, BA, CFA, ACCA Esq. @ParikPatelCFA]

A guide to living in New York City

[Timo Witten @TimoWitten]

@TrackingActions @lufthansa Beautiful!

[Timo Witten @TimoWitten]

Be cautious when using deep learning to study the brain

[bioRxiv Neuroscience @biorxiv_neursci]

No Free Lunch from Deep Learning in Neuroscience: A Case Study through Models of the Entorhinal-Hippocampal Circuit https://t.co/8ZtMqoXVmD #biorxiv_neursci

[Timo Witten @TimoWitten]

@SaraASolla @tyrell_turing @TonyZador @KordingLab @recursus @ShahabBakht @NXBhattasali @joshdubnau @WiringTheBrain @L_andreae Thanks, Sara! It is so important to have you here to tell the truth. Some people on Twitter with many followers make wrong statements that can mislead young students.

[Oded Rechavi @OdedRechavi]

Almost every abstract has a line that says "however, bla bla bla is still poorly understood" and it's never clear what's the cutoff for being understood

[Timo Witten @TimoWitten]

If she is ok, I’m ok.

[Fermat's Library @fermatslibrary]

During the 17th century, this deceptively simple series, eluded the efforts of many mathematicians, including Newton, Liebniz and Jakob Bernoulli. It was Euler that in 1734, at the age of 27, derived the solution: π²/6

[Timo Witten @TimoWitten]

@KordingLab @somnirons Almost everything can approximate the backpropagation algorithm.

[Gabriel Peyré @gabrielpeyre]

Optimal quantization solves a minimum Optimal Transport projection. Lloyd algorithm computes a stationary point of this non-convex energy (see work by Canas and Rosasco https://t.co/yladoE9Q4O)

[Heidi Seibold @[email protected] @HeidiBaya]

I asked participants in a workshop, what is stopping them from being an #OpenScience champion. This is what they responded.

Is it worrying that they put "system" at the top?

Anything missing here in your opinion?

[Timo Witten @TimoWitten]

@behrenstimb @TPVogels I was joking. You greatly contributed to the community by setting @worldwideneuro up during the pandemic. I am also looking forward to seeing more of your research work.

[Timo Witten @TimoWitten]

@behrenstimb @TPVogels I agree with the first part.