Balázs Kégl @balazskegl Twitter profile

Pinned Tweet

Balázs Kégl

2 days

If you like @vervaeke_john @PageauJonathan @PaulVanderKlay like I do but you find them too gnostic (bodiless talking heads), like I do, your natural next stop is @Bardissimo . Check out this excellent conversation with @andrewpgsweeny

Sweeny vs Bard Season 2 Ep. 12: Jordan Peterson, the trickster, and...

NEW PODCAST VERSION: https://anchor.fm/andrew-sweenySupport this channel and my projects by subscribing or helping on Patreon:Patreon: https://www.patreon.c...

www.youtube.com

0

4

Last Seen Profiles

@RylandPenn93130

@bokeplokalmalam

@majidd3m

@nangfa06574395

@Natas_senkai

@arevalooand

@phet11225

@bokeplokalmalam

@Brett_Costello

@_Obscur8_

@Mononi15

@TXStateTrack

@AzegamiTam55568

@bokeplokalmalam

@aprisiarene

@Dorian58990295

@bokeplokalmalam

@l3osskiss

@xxky0o

@jandakembangstw

@Its_Lira

@penikma89063156

@ipacell3939

@OfficialAhee

@theXblueprint

@lawyer54344

@lhsingapura

@Ash_TheK

@bokeplokalmalam

@ROA_YOGU

@covill_mo

@SgtPeteSimpson

@lefantomee_

@analucialds

@galery_basah10

@ykcobey

Balázs Kégl

@balazskegl

3 years

I feel angry that it is so hard to publish machine learning contributions on small data. Most problems are not images and NLP, and most data scientist work on relatively small tabular data. And most of the _value_ is there. And we know very little about that regime.

27

86

707

Balázs Kégl

@balazskegl

3 years

What do you think about this argument? GPT-3 is obviously not intelligent. It was fed by a nontrivial percentage of all text produced by humans which cannot grow by several orders of magnitude. Therefore the paradigm of feeding data into function learning is a dead end for AGI.

57

31

418

Balázs Kégl

@balazskegl

9 months

@ylecun Thanks Yann, I think your political role now is as important as your technical role.

4

3

228

Balázs Kégl

@balazskegl

7 months

@ylecun I don't think it's that simple. The most awesome property of living cognition is not the absorption of 1E15 bytes of video "data" the child sees but its filtering. How little it actually needs to "see" to build a world model. And the key is that a child is not a mere observer,

7

100

Balázs Kégl

@balazskegl

7 months

@ylecun You can only learn causal models by embodying the AI and let it manipulate the world, running experiments. Watching videos and reading text will not get you to causal world models, whatever the architecture and the task you define.

A call for embodied AI

We propose Embodied AI as the next fundamental step in the pursuit of Artificial General Intelligence, juxtaposing it against current AI advancements, particularly Large Language Models. We...

arxiv.org

10

2

85

Balázs Kégl

@balazskegl

4 years

We are hiring at Huawei's Paris AI research! - PhD in 2020 or 2021 - permission to work in France - strong publication record - preferable domains: RL, bandits, physical systems, time series, AutoML, combinatorial and online optimization, transfer learning RT please

0

21

60

Balázs Kégl

@balazskegl

7 years

Gender balance at a #NIPS2017 workshop panel, sooo refreshing #nips4creativity #creativeAI @douglas_eck

1

14

59

Balázs Kégl

@balazskegl

8 years

The death of lecture. Why on earth should I prepare a course on #convnets other than a Q&A on this?Thanks @karpathy

0

8

60

Balázs Kégl

@balazskegl

10 months

@ylecun @EduardoSlonski @DrJimFan @RichardSSutton We also have 10s to 100s millions intero and exteroceptive neurons (estimate by GPT :) ). Also, we don't just given a set of data and learn a function, we live, the paradigm is way more complex than anything we do in ML. The rest of the body, may be as important as the network

4

1

56

Balázs Kégl

@balazskegl

1 year

@martinmbauer But I love the supercoductor discussion on twitter. It's a new category, halfway between poetry and science.

0

51

Balázs Kégl

@balazskegl

7 years

My slides: A historical introduction to #DeepLearning : hardware, data, and tricks #bigdata #datascience #ML

0

36

50

Balázs Kégl

@balazskegl

10 months

I think "Bayesian" is mostly a synonym of "subjective", probabilities unsupported by third-person scientific data. I think there is nothing wrong with quantifying feelings and even asking for validation, just let's be clear about what we're doing.

Christoph Molnar

@ChristophMolnar

10 months

Two types of Bayesians 1) Beer pub Bayesian: People who want to be seen as intelligent by using words like "Bayesian", "prior", "posterior", "updating" to refer to the fact that people have opinions and sometimes change their minds. 2) Actual Bayesian who analyzes data

22

30

365

5

7

49

Balázs Kégl

@balazskegl

3 years

@davidasboth The surprise of the last year for me that you can train neural nets on 200 points. OK, it's smooth low-dim data. But you _do_ need NNs, linear or forests do not work. And the experiments generated strange puzzles over heteroscedastic NN regression.

Model-based micro-data reinforcement learning: what are the crucial...

We contribute to micro-data model-based reinforcement learning (MBRL) by rigorously comparing popular generative models using a fixed (random shooting) control agent. We find that on an environment...

openreview.net

2

8

48

Balázs Kégl

@balazskegl

11 months

@tszzl Yes. Banks seems to overlook the spiritual way though that we also have a tendency to follow facing technocratization. Culture is changing, but so do we, individuals. With food and shelter assured, more and more people will have time to explore enlightenment, which will

9

5

43

Balázs Kégl

@balazskegl

3 years

A constructive idea: let's have a #5000 track or tag at conferences, in which papers only use training data with <= 5000 instances.

3

1

43

Balázs Kégl

@balazskegl

2 years

. @vervaeke_john proposes here an alternative to the Turing test wherein we let two copies of the subject to talk to each other. His scientific prediction is that within a minute the conversation, and I quote, "goes batshit crazy". I tested GPT-3. 👇

✂️ Batshit crazy

45 seconds · Clipped by Balazs Kegl · Original video "John Vervaeke and Paul Vander Klay Dialogos on Consciousness" by Paul VanderKlay

www.youtube.com

1

10

42

Balázs Kégl

@balazskegl

7 months

@GaryMarcus Given the enormous, unimaginable suffering in the 20th century, should we have stopped the Enlightenment? Had Luther foreseen the death of God, would he have gone ahead with his proposals?

4

0

42

Balázs Kégl

@balazskegl

2 years

I've been in this ML business for about 30 years now, and still can't get around the fact that the optimal random forest overfits like crazy. This is not how the world supposed to be.

7

3

39

Balázs Kégl

@balazskegl

6 years

A major milestone: our #RAMP preprint paper is out. Don't hesitate to comment or discuss our vision on the openreview site if you are interested in #transparency and #reproducibility in scientific #datascience workflows.

1

26

40

Balázs Kégl

@balazskegl

7 months

@ylecun True. What you fundamentally cannot learn by mere observation is what data you should learn from. What prediction is important/useful to care about. We are trying to "teach" AI everything to compensate this lack, and quantity covers it up superficially. But this is the main

2

38

Balázs Kégl

@balazskegl

6 years

We have concluded the #ASDchallenge of @SaclayCDS @institutpasteur @IESFfrance , 123 participants, 677 submissions, AUC moved from 0.68 to 0.81. Congrats to the winners and welcome to the collaborative phase! #datascience for #healthcare #MachineLearning

2

22

38

Balázs Kégl

@balazskegl

1 year

I love that ICML accepts these kinds of papers, actually it's pretty surprising, we didn't even think about aiming. Just look at this plot and promise me never again using standard error as your confidence estimate on heavy-tail RL losses. Andrew Patterson, Samuel Neumann,

4

7

34

Balázs Kégl

@balazskegl

3 years

Anybody else noticed that in reinforcement learning, performance across seeds is non-Gaussian? How do you cope with it? This means that reporting mean and even standard deviation is not that meaningful, we're averaging qualitatively different things.

6

1

34

Balázs Kégl

@balazskegl

2 years

I don't know about "batshit crazy", but as much as I can tell, it is the right scientific term. More precisely, it's like Sheldon trying to pick up a woman in a bar.

3

4

35

Balázs Kégl

@balazskegl

3 years

Let's celebrate the 30th anniversary of Rich Sutton's Dyna paper. Four pages, zero experiments, but one full page on limitations, 600 citations, and the founding of a sub-branch in reinforcement learning on which I am betting for AGI.

0

2

35

Balázs Kégl

@balazskegl

2 years

Forget ChatGPT, here is what's coming. As all devices will be equipped with self-optimizing AIs, they will naturally develop their personalities (since their experiences will differ). This means that engineering will go from designing "machines" to "shepherding" little AIs.

6

4

30

Balázs Kégl

@balazskegl

7 years

As long as it lives in a perfectly simulated world with a known and well defined objective.

The Economist

@TheEconomist

7 years

The latest AI can work things out without being taught

12

229

279

1

13

33

Balázs Kégl

@balazskegl

10 months

@roydanroy It's how you feel after someone defines "understanding" for you.

1

2

30

Balázs Kégl

@balazskegl

10 months

@ProfNoahGian Yes. Your wife would say that this is the nerd's way of asking for validation of our feelings.

Balázs Kégl

@balazskegl

10 months

I think "Bayesian" is mostly a synonym of "subjective", probabilities unsupported by third-person scientific data. I think there is nothing wrong with quantifying feelings and even asking for validation, just let's be clear about what we're doing.

5

7

49

1

3

30

Balázs Kégl

@balazskegl

9 months

@QiaochuYuan I like him, like his approach, and happy that he has success. He inspired me to start my podcast. You don't need to listen to him. You can also do what he does, follow your own instincts and manifest the way you think conversations should be done. It's enlightening.

1

0

27

Balázs Kégl

@balazskegl

1 year

The difference between planning in #reinforcementlearning and in #LLM is that in RL we check where we really are after each action, and replan. This means that the agent needs to be connected to the world. This rhymes with what @ylecun has been arguing.

#3 Bogdan Cirstea: GPT and AI safety

00:00:00 Intro. Bogdan's journey through the French system towards PhD in AI. Inspiration by early DeepMind papers, research on LSTM and other recurrent arch...

www.youtube.com

4

3

24

Balázs Kégl

@balazskegl

2 years

Catchy title (GPT-3 is not a god to be believed in), but the article is hyper-precise. I think GPT-3 is a nice tool but missing grounding. Also nice to see how my vision articulates with @ylecun 's. IA : Pourquoi Huawei ne croit pas à GPT-3 via @zdnetfr

IA : Pourquoi Huawei ne croit pas à GPT-3 - ZDNET

Technologie : Au salon international sur l'intelligence artificielle, qui s'est tenu à Cannes la semaine dernière, Balazs Kegl a levé le voile pour ZDNet sur son travail de recherche au sein du...

www.zdnet.fr

0

2

24

Balázs Kégl

@balazskegl

6 years

If you are a #datascientist in the #Paris area and would like to work on cool (honestly unprecedented) #healthcare data and maybe win some prizes, consider joining the #MachineLearning #ASDchallenge . @SaclayCDS @institutpasteur @IESFfrance @ParisMLgroup

IMaging-PsychiAtry Challenge: predicting autism

A data challenge on Autism Spectrum Disorder detection What is this challenge? Autism spectrum disorder (ASD) is a severe psychiatric disorder that affects 1 in 166 children. There is evidence that...

indico.ijclab.in2p3.fr

0

16

24

Balázs Kégl

@balazskegl

9 years

The #datascience ecosystem: actors, incentives, challenges. @SaclayCDS @U_ParisSaclay http://t.co/6tzYefS0D0

0

23

Balázs Kégl

@balazskegl

8 years

Tired of ad optimization? Work w us @SaclayCDS on #DataScience in astro, health, ecolo w @agramfort @GaelVaroquaux

0

19

23

Balázs Kégl

@balazskegl

2 years

As we become less tribal, a major source of conflict is perceptual diversity. That we cannot imagine that others perceive the world differently from us. @anilkseth

A key to consciousness could lie in “perceptual diversity”

Every brain experiences reality differently. Vox spoke to neuroscientist Anil Seth about the Perception Census.

www.vox.com

0

5

23

Balázs Kégl

@balazskegl

2 years

This should be the default condition to write NeurIPS papers.

3

0

21

Balázs Kégl

@balazskegl

10 months

"The immune system of a pregnant person is mindblowing" Amazing paper on how selves can be nested by @AnnaCiaunica @drmichaellevin @_fernando_rosas and Karl Friston Watch @AnnaCiaunica explaining the gist:

#7 Anna Ciaunica: embodied cognition and depersonalization

00:00:00 Intro: empirical vs armchair philosopher. Visual vs tactile understanding.00:06:08 How subjective experience rises from physical matter. Entering wi...

www.youtube.com

Anna Ciaunica @annaciaunica.bsky.social

@AnnaCiaunica

10 months

Nested Selves: Self-Organisation and Shared Markov Blankets in Prenatal Development in Humans via @OSFramework

2

16

65

1

8

22

Balázs Kégl

@balazskegl

3 years

Well, I think that's exactly the point of @GaryMarcus . Images that fool us need to be carefully engineered, and play no role in our survival (on the contrary, our "illusions" are probably necessary for survival), whereas deep nets have an arbitrary relevance landscape.

Yann LeCun

@ylecun

3 years

@GaryMarcus @savvyRL @scaleai @allen_ai @anh_ng8 @jeffclune @jasonyo Happy birthday human neural nets! You are nearly 1,000,000 years old now, and still utterly fooled by images like these.

11

24

221

2

3

22

Balázs Kégl

@balazskegl

9 years

. @agramfort is explaining his super cool fast Lasso at #icml2015 . Soon in @scikit_learn . http://t.co/ABfqlxfNyi

1

18

21

Balázs Kégl

@balazskegl

8 years

Looking outside the #DataScience bubble. @lawrennd on the issues of sampling bias and data valuation.

0

5

22

Balázs Kégl

@balazskegl

11 months

I agree with @drmichaellevin that we have been producing high-level agential intelligences aka kids, without asking for guarantees about their alignment. That raises the following poll. Retweet if you're interested in the answer. Do you have

kids, worried AI alignmnt

55

kids, not worried

105

no kids, worried

64

no kids, not worried

209

18

8

22

Balázs Kégl

@balazskegl

6 years

My slides for the #DataScience colloquium @ENS_ULM today at noon in Jaurès Room (29 rue d’Ulm). Thanks @KrzakalaF @gabrielpeyre for the opportunity. #machinelearning #datachellenge #design

2

9

21

Balázs Kégl

@balazskegl

8 years

Why do some developers at strong companies like Google consider Agile development to be nonsense? by David Jeske

0

13

21

Balázs Kégl

@balazskegl

6 years

"Overall, there is a huge mismatch in what ML researchers consider “sexy” and what is important for ML practitioners!"

0

6

20

Balázs Kégl

@balazskegl

2 years

@BeingMIAkashs @ylecun Model-predictive control on a learned model _is_ reinforcement learning. It is called model-based reinforcement learning. This is like saying: "Abandon Machine Learning, in favor of optimization".

2

5

20

Balázs Kégl

@balazskegl

3 years

@OkbaLeftHanded That's true but in my experience e.g. small data reinforcement learning has a huge market, trillions per year. It's not the IT sector, more heavy (control) engineering and transportation so it's not the usual high-liquidity companies that dominate ML research today. But it's big.

2

20

Balázs Kégl

@balazskegl

2 years

@moreisdifferent @tdietterich Yes, we have the same experience in model-based reinforcement learning. I hate it because it's dumbly inefficient, but retrain from scratch on the full data each time a new data arrives is unbeatable for now.

1

0

20

Balázs Kégl

@balazskegl

10 months

If you find Sapolsky's no free will thesis appealing, read this dialog. Performative contradictions are best demonstrated through stories.

Michael Levin

@drmichaellevin

10 months

A short dialog between an applicant who doesn’t believe in free will and a hiring manager at a software company:

27

19

101

3

5

18

Balázs Kégl

@balazskegl

11 months

@casebash @drmichaellevin Best to listen to @vervaeke_john . We don't align kids by giving them and enforcing rules (purely), but by embodied stories and rituals, and by bringing them in contact both with others and with reality, so they bind to both.

AI: The Coming Thresholds and The Path We Must Take | Internationally...

Dr. John Vervaeke lays out a multifaceted argument discussing the potential uses, thresholds, and calamities that may occur due to the increase in artificial...

www.youtube.com

1

3

19

Balázs Kégl

@balazskegl

6 years

On the ambitions of @scikit_learn , @GaelVaroquaux at the inauguration of the scikit-learn consortium @StationF @Inria

0

5

18

Balázs Kégl

@balazskegl

6 years

The @SaclayCDS Solar wind #RAMP event of tomorrow is open. The goal is to detect variable-length events in multiple aligned time series. Don't hesitate to sign up if you would like to participate tomorrow or in the next two weeks.

0

10

17

Balázs Kégl

@balazskegl

1 year

@CRSTNNSL @DavidDeutschOxf I think it may be purity vs mixing. There was less population density, languages developed "on their own" for 50K years. Then with the Indo-European explosion, there was much more mixing, and mixing meant to get rid of fragile complexities. You can see it in modenr-day English,

3

1

17

Balázs Kégl

@balazskegl

3 years

First of the three lectures on model-based reinforcement learning/self-driving systems at the DeepLearn summer school, Gran Canaria 2021.

Model-based reinforcement learning/self-driving systems 1 DeepLearn...

First of my three-lecture series on model-based micro-data iterative batch reinforcement learning for optimizing engineering systems.https://www.slideshare.n...

www.youtube.com

1

5

18

Balázs Kégl

@balazskegl

7 years

My slides for today's #NIPS2017 workshop talk on collaborative #datachallenge and #datascience process management. Room S1 in ten minutes.

0

10

17

Balázs Kégl

@balazskegl

9 years

My slides on "What's wrong with data challenges" at tomorrow's #NIPS2015 ALEPH WS #datascience #OpenInnovation

What is wrong with data challenges

What is wrong with data challenges - Download as a PDF or view online for free

www.slideshare.net

0

8

17

Balázs Kégl

@balazskegl

9 months

LOL, techcrunch just called @scikit_learn , XGBoost and the full data science stack "good old fashioned AI". I don't know who'll be more upset, @GaryMarcus or @tunguz and @GaelVaroquaux ?

Good old-fashioned AI remains viable in spite of the rise of LLMs | TechCrunch

There is clearly still a role for task-specific models because they can be smaller, faster, cheaper, and in some cases be more performant.

techcrunch.com

5

0

16

Balázs Kégl

@balazskegl

1 year

Happy Eastern!! Goulash a la cowboy.

2

1

17

Balázs Kégl

@balazskegl

7 months

@ylecun Yes, it's the same category error: LLMs still cannot bike. Propositional and procedural knowing is not the same.

0

3

16

Balázs Kégl

@balazskegl

4 years

On the way to AGI, I think pure model-free RL is a dead end. Yes it is difficult for an agent to figure out what slice of the world matters for reaching its goals but this is not an excuse for not learning a model for that slice. Every living thing builds such a model. Pls refute

4

0

17

Balázs Kégl

@balazskegl

11 months

My conversation with @AnnaCiaunica about embodied cognition, feeling like a zombie, depersonalization, meditation, movement therapy, 1st person science, and much more. Pls give it thumbs up, share, comment, and sign up!

#7 Anna Ciaunica: embodied cognition and depersonalization

00:00:00 Intro: empirical vs armchair philosopher. Visual vs tactile understanding.00:06:08 How subjective experience rises from physical matter. Entering wi...

www.youtube.com

2

3

16

Balázs Kégl

@balazskegl

10 months

@TonyZador Hm, you put the genome in a test tube, it won't hunt. You put the colt in water, it won't stand. The point is that there are a lot of environmental and developmental affordances and constraints that produce the behavior, and those are also information.

2

0

14

Balázs Kégl

@balazskegl

1 year

@fchollet That's what we use in Hungary.

0

16

Balázs Kégl

@balazskegl

9 months

This is way scarier than AGI.

Sheel Mohnot

@pitdesi

9 months

I am often surprised by polling but this one scares the shit out of me. 1 out of 5 people 18-29 believe the holocaust is a myth. How did it get to be this way?

944

619

5K

0

3

16

Balázs Kégl

@balazskegl

7 months

Preprint "A call for embodied AI" with my coathors @_GPaolo and Jonas Gonzales-Billandon.

A call for embodied AI

We propose Embodied AI as the next fundamental step in the pursuit of Artificial General Intelligence, juxtaposing it against current AI advancements, particularly Large Language Models. We...

arxiv.org

1

5

14

Balázs Kégl

@balazskegl

7 years

Insect recognition #datachallenge w 5K€ prizes. #Datascience #DeepLearning #Keras #PyTorch @SaclayCDS @IESFfrance

0

13

16

Balázs Kégl

@balazskegl

7 years

Congratulations to my and @AKAZAKCI 's PhD student @mehdidc defending his thesis on #DeepLearning and novelty generation. If you want to understand why we count letters generated by a model trained on digits when we don't even want to generate letters, ask him a copy of his thesis

0

1

16

Balázs Kégl

@balazskegl

9 years

LSTM in ten lines of code #tensorflow #NIPS2015

0

4

15

Balázs Kégl

@balazskegl

10 months

DALLE is definitely out of sync :) "Realistic photo of a young tourist-looking guy smuggling weights of a large language model in his backpack at US airport customs in Chicago."

6

8

16

Balázs Kégl

@balazskegl

2 years

@bernhardsson I have an upcoming preprint (brutally downscored by NeurIPS :) ) that shows that some of the Ray Tune engines beat random search by a factor of 2x to 3x. Best ones seem AX, BlendSearch, Hebo.

0

16

Balázs Kégl

@balazskegl

1 year

OK, I've seen stuff. But 6000 French CEOs celebrating my 17-year-old son with a standing ovation still feels surreal. What a day! I'm immensely proud, bravo Martin! Thank you @tweets_Apm !

Apm

@tweets_Apm

1 year

Martin a construit le coffre #Apm2023 , la fameuse capsule temporelle qui sera ouverte en 2050 !

1

6

2

0

15

Balázs Kégl

@balazskegl

11 months

@DrJimFan @tegmark No debate, I agree :). What I would add is that "world" is a relative concept. It's what we sense. So LLM has implicit information about our world model. A WiFi antenna's world model will be quite different. In that sense LLMs will be helpful for an agent whose world model

0

1

15

Balázs Kégl

@balazskegl

7 years

How to build a #datascience pipeline. Start with the prediction target y. Involve #IT early.

1

9

14

Balázs Kégl

@balazskegl

6 years

The @SaclayCDS @institutpasteur @IESFfrance launch the Autism Spectrum Disorder #datachallenge . The competition will end on July 1st, followed by a collaborative phase. We award 9.5K€ in money prizes to the top ten participants. #ASDchallenge #datascience

0

14

15

Balázs Kégl

@balazskegl

7 years

My slides on #ramp workflow and kits for today's @ParisMLgroup meetup. Build your #datachallenge . Organize teamwork.

0

7

15

Balázs Kégl

@balazskegl

7 years

Teaching the #datascience process. #machinelearning #flippedclassroom @Polytechnique @SaclayCDS

0

11

15

Balázs Kégl

@balazskegl

11 months

My conversation with @GaelVaroquaux , founder of @scikit_learn , on #MachineLearning , #healthcare , the importance of communicating why scientists do what they do. Enjoy! Pls share, comment, like, and sign up to the I, Scientist podcast.

#5 Gaël Varoquaux: scikit-learn, ML in health and social science,...

00:00:00 Intro. Gaels's journey from physics to AI through coding, then health and social science applications.00:06:17 Looking for impact. Are we using our ...

www.youtube.com

1

5

14

Balázs Kégl

@balazskegl

4 years

The four reviews we got on our NeurIPS paper on model-based RL are excellent. The paper will be rejected but it's rather because of systemic issues and the lack of a clear scientific paradigm in ML than any misreading or not reading of the paper by the reviewers.

2

3

15

Balázs Kégl

@balazskegl

4 years

@vgr I built a radiator in my bedroom out of the cooling grid of an old fridge, a plastic pipe, and some corks during the 97 ice storm in Montreal. Electricity was out for a week but gas water heater worked in the bathroom.

0

2

13

Balázs Kégl

@balazskegl

2 years

@Aella_Girl I'm with David Deida on this. That desire is like the air I'm breathing, life energy that I can turn into creation. No shame to have it and no suppression. But I don't act on it because I'm monogamous and what we're building with my partner could not exist if we were poly.

1

0

13

Balázs Kégl

@balazskegl

11 months

@sentientist Think about them this way: you are their frontal lobes. They are born immature. So their behavior is a mirror of your toddler self. My worst rage is coming when they do something I was not allowed to. My inner toddler shrieks, IT'S NOT FAIR!!! Golden material to bring into

0

14

Balázs Kégl

@balazskegl

7 years

How to build a #datascience pipeline. Preparing my talk at #OReillyAI in the #AI @xprize @ibmwatson workshop.

0

14

Balázs Kégl

@balazskegl

7 years

We launched our first "Kaggle RAMP": a collaborative RAMP team to compete in a @kaggle challenge. Feel free to join!

0

9

13

Balázs Kégl

@balazskegl

2 years

Entropy is ~ the log number of microscopic states that produces the same macroscopic state, right? So can we say that without postulating _levels_ (at least a micro and macroscopic level of analysis), entropy loses its grounding?

1

14

Balázs Kégl

@balazskegl

8 years

A short note on reforming #machinelearning #conference reviewing. Let's make the annual #ICML / #NIPS lottery...

Embrace the random

A case for randomizing acceptance of borderline papers

balazskegl.medium.com

2

7

12

Balázs Kégl

@balazskegl

1 year

I'm super excited to announce my podcast: I, scientist, where we explore AI, body, and soul. My first guest is @_GPaolo whom we discuss agency in AI, reinforcement learning, and much more. Pls give it thumbs up, subscribe, comment, and share.

2

14

Balázs Kégl

@balazskegl

6 years

Our next single-day @SaclayCDS #RAMP hackathon will be on October 10 at the INRIA Turing building on detecting Solar storms from spacecraft measurements (event detection in time series). If you would like to participate, please sign up before October 3:

RAMP 12: Solar storm detection

The goal of this RAMP is to detect Interplanetary Coronal Mass Ejections (ICMEs) in the data measured by in-situ spacecraft. ICMEs are the interplanetary counterpart of Coronal Mass Ejections (CMEs),...

indico.ijclab.in2p3.fr

0

15

13

Balázs Kégl

@balazskegl

8 years

Title page of my #nips2016 WS talk tomorrow in room 127-128 at 14h45. Title is written by a machine trained only on digits.

0

3

14

Balázs Kégl

@balazskegl

7 years

The first ever scientific paper where optimization of the #datascience analysis pipeline was entirely #crowdsourced .

0

15

14

Balázs Kégl

@balazskegl

7 years

"Elements of causal inference" by Jonas Peters, Dominik Janzing and @bschoelkopf is out, #openaccess , at the top of my reading list.

0

1

14

Balázs Kégl

@balazskegl

8 years

Breaking news

2

25

14

Balázs Kégl

@balazskegl

2 years

@hardmaru @slashML It's painstakingly slow because there are so many moving parts and they all depend on each other. Also takes a couple of years to build a good experimental stack. Third: benchmarks are all over the place, experimental settings are not fixed.

2

0

14

Balázs Kégl

@balazskegl

6 months

Are we rushing ahead into the unknown, scaring people, while also producing fake shelters for them to retreat into artificial small worlds? Are we losing our grip on the world or realizing our humanness in the mirror of AI? Will AI take our agency away? Are we willingly giving it

1

3

14

Balázs Kégl

@balazskegl

7 years

My talk on collaborative #datachallenges and managing the #datascience process @franceisai

0

6

13

Balázs Kégl

@balazskegl

6 years

The @SaclayCDS is proudly serving the backend of an insect recognition Android app developed by Leonard Boussioux. @Le_Museum #DataScience #MachineLearning #citizenscience #crowdsourcing #ecology

31K views · 172 reactions | CGD2 : Léonard Boussioux présente...

Ils sont étudiants et ont des idées pour innover dans le développement durable. Découvrez les lauréats du Concours Génération Développement Durable ! 🌱...

www.facebook.com

0

4

13

Balázs Kégl

@balazskegl

4 years

Our #ICLR2021 paper on model-based RL with @albertcthomas and Gabriel Hurtado Blog: We compare generative models on Acrobot. Both multimodality and heteroscedasticity are crucial, even though the environment is noiseless. 1/n

Model-based micro-data reinforcement learning:

What are the crucial model properties and which model to choose?

towardsdatascience.com

1

3

12

Balázs Kégl

@balazskegl

7 months

We've just submitted a position paper on embodied AI. One of my realizations during writing is that social media AI is more embodied than LLMs, making them more natural to interact with, which may explain why there is less buzz about their alignment, even though their impact on

1

3

13

Balázs Kégl

@balazskegl

1 year

@fchollet You underestimate the reliability of the source when we judge information. Our whole brain is about judging epistemic authority. I can't check the veracity of every piece of information, it defeats the purpose of distributed intelligence.

1

0

13

Balázs Kégl

@balazskegl

2 years

@Philip_Goff Can you elaborate how theists deny suffering? My understanding would be more like they make suffering meaningful, thus bearable. Avoiding suffering is not at the top of the value hierarchy, but suffering exists.

1

0

12

Balázs Kégl

@balazskegl

1 year

@hubermanlab @lexfridman For what would he give the yearly Fridman prize and why?

0

13