Ulrich Junker

Scientist in Artificial Intelligence & Decision Sciences.
🚨Escaped to

[Ulrich Junker @UlrichJunker]

From Nov 4, 2022 on, I will use this account only occasionally. Please follow me at @[email protected] on @joinmastodon for spontaneous discussions and sharing of information.

[ECAI2025 @ecai2025]

From yesterday, to close the 2nd ECAI day, nothing better than a new #MeetECAIers at the WCCC:

😂🗣️ “ So whenever you come into my office and you discover me playing a game, it's research “ - Jonathan Schaeffer

#ECAI #AI

[ECAI2025 @ecai2025]

Insightful session by @IrynaGurevych on tackling misinformation with NLP and multimodal AI! 📚🔍 From debunking fake science to analyzing manipulated visuals. #Misinformation #AI #NLP

[Preference Handling @AdvancedPref]

#MPREF2024 has been held successfully. Our thanks go to
🖊️The PC for high quality reviews.
🚀 @ECAI2024 for providing us with the infrastructure.
⭐️ Jörg Rothe & Khaled Belahcene for their invited talks.
🗞️ The authors for their interesting submissions.
https://t.co/2THDK0Zopv

[Subbarao Kambhampati (కంభంపాటి సుబ్బారావు) @rao2z]

So one of my former students, Yuheng, asked that great Panjandrum--ChatGPT--to shed light on this year's Physics Nobel..😂🤣

[Noam Brown @polynoamial]

Today, I’m excited to share with you all the fruit of our effort at @OpenAI to create AI models capable of truly general reasoning: OpenAI's new o1 model series! (aka 🍓) Let me explain 🧵 1/

[Ulrich Junker @UlrichJunker]

[Solved] Where do the rumors about the dog eating come from?
- see video
https://t.co/DjgKPYEjBZ

[Ulrich Junker @UlrichJunker]

Sehr gutes und Boden-berührendes Webinar zur kommenden Thüringenwahl. #Klimaschutz kein Thema?

[Europe Calling e.V. @_europecalling]

Heute Abend um 20 Uhr schauen wir mit @Luisamneubauer , Prof‘in Cornelia Betsch @IPB_Erfurt und Bernhard @stengele2024 nach Thüringen!

[Preference Handling @AdvancedPref]

#MPREF2024 program is now online! Please join us on Oct 20 in Santiago de Compostela for an exciting event on computational methods for #PreferenceHandling.

#ECAI2024 workshop

⚠️ Early Registration Deadline of ECAI: August 19, 2024

https://t.co/2THDK0Zopv

[Ulrich Junker @UlrichJunker]

OTD 20 years ago, I presented #QuickXplain at #AAAI2004. Now it is time to tell the truth about its discovery.
https://t.co/0EH5E1VkKr

[Ulrich Junker @UlrichJunker]

@Anabelle8881986 Digital identity transformation! I was looking for a tweet by Pierre Schaus, but this tweer is lost for humanity. May be we can find it in some LLM’s training data …

[Ulrich Junker @UlrichJunker]

Ich habe gerade hier unterschrieben: Der Klimaschutz in Europa muss weitergehen. Von der Leyen darf den Green Deal nicht aufgeben!

Hier kannst auch du unterschreiben und aktiv werden: https://t.co/hyY4vktdtA

[Ulrich Junker @UlrichJunker]

I have just signed the petition "Urgent: Your voice for nature". You can also sign it here:

https://t.co/6yWXyFNQ9m

#SaveOurNature

[Ulrich Junker @UlrichJunker]

@rao2z We should not ask to stop anthropomorphizing LLMs, but to start de-personifying LLMs!

Similar to this: “Don’t take your puppet for a real person, kid! It won’t replace your friends.”

What about: “Don’t take this chatbot for a real person, boss! It won’t replace your employees.”

[Ulrich Junker @UlrichJunker]

@rao2z #Anthropomorphizing is a quite complicated word and it suggests that people spent effort to attribute human-like qualities to #LLMs. In reality, they are just lacking #CriticalThinking and imagination. When reading text, they automatically attribute it to a human personality. ⤵️

[Subbarao Kambhampati (కంభంపాటి సుబ్బారావు) @rao2z]

Why we should view LLMs as powerful Cognitive Orthotics rather than alternatives for human intelligence #SundayHarangue

LLMs are amazing giant external non-veridical memories that can serve as powerful cognitive orthotics for us, if rightly used (c.f. https://t.co/rvG2e5tlfa).

The trick, IMHO, is to exploit them without deluding ourselves in the process.

The delusion comes chiefly from our incessant need to confuse them for human intelligence, merrily applying anthropomorphic concepts such as thinking , thoughts, reasoning and self-critiquing to LLMs. (c.f. https://t.co/MPUdvj3AwC;
https://t.co/q93JRiE7XS;
https://t.co/vUXwazqwv9)

This anthropomorphization is quite futile--and, as shown in the case of some of the current Ersatz Natural Science AI literature--even counterproductive and misleading.

Sure we didn't quite foresee how impressive the approximate omniscience of these n-gram models on steroids would be, but that doesn't have to make us assume they do everything humans do.

Unless human-level #AI is your singular goal, you don't necessarily need to think auto-regressive LLMs suck (as @ylecun puts it colorfully).

LLMs can be very effective complementary cognitive orthotics without subsuming human intelligence.

LLMs do some things way way way better than humans do  (the litmus test--from my perspective--being converting anything to iambic pentameter in seconds 😋https://t.co/OvFDYveC71), and do other things (planning, reasoning, self critiquing, mental modeling) much worse (c.f. ;
https://t.co/F7DX0DIRE6;
https://t.co/BRSU4AyGji; https://t.co/DS1PitXIdP)

If we can manage to tone down the "LLMs are Zero-shot <XXX>" studies rife with confirmation biases that conflate approximate retrieval for other capabilities associated with human intelligence (c.f. https://t.co/Sz35rUmr7j; https://t.co/XI5aQRxXYO), we can:

1. Focus on the right way of leveraging the strengths of LLMs. This can certainly be done in LLM-modulo architectures, with either humans or other specialized sound reasoners in the loop).

2. To the extent human-level intelligence is still your holy grail, keep open research avenues that don't just involve scaling up autoregressive architectures. (It is in this sense that I sympathize with @ylecun's AR-LLMs suck comment--they do, currently, suck the oxygen out of the research room 😅)

tldr; you can get much farther with LLMs & #AI if you think of them as cognitive orthotics and can stop anthropomorphizing them.. (c.f. https://t.co/BFlV0jUkIG;
https://t.co/8XgJZn6CLv)

[Jutta Paulus @JuttaPaulusRLP]

Die #Energiewende zu beschleunigen, macht sie billiger.

Leute, das sagen nicht "grüne Spinner."
Das ist die .@IEA, die Internationale Energie-Agentur. Gegründet in der Ölkrise der 70er. Für Zahlen, Daten, Fakten über alle Energieträger weltweit.

Worauf warten?
#VoteForClimate

[Jutta Paulus @JuttaPaulusRLP]

Morgen!

Diese Europawahl ist eine Richtungswahl.
Weiter mit #GreenDeal, #Klimaschutz, Nachhaltigkeit?
Oder Rollback, Umweltschutz abbauen und Zusammenarbeit mit Rechtspopulisten?

Jede Stimme zählt!
#VoteGreen

[Ulrich Junker @UlrichJunker]

@marlieskorsten1 @JuttaPaulusRLP Informieren Sie sich über Klimawissenschaft. Das ganze ist kein Spiel. Das ist bitterer Ernst.

[Subbarao Kambhampati (కంభంపాటి సుబ్బారావు) @rao2z]

📢📢 Announcing the availability of our monograph on "Explainable Human-AI Interaction," co-authored with @sarath_ssreedh & @anaghapk (with the astral background presence of @tchakra2) on your neighborhood @arxiv..

Read, share, and build on it! 1/

👉https://t.co/lEn82AwpfL