iiinova — where your digital self keeps evolving

Reposted by Pasquale Minervini

EACL 2027 @eaclmeeting.bsky.social · 17/08/2026

The CFP for the Student Research Workshop at #EACL2027 is up! 🔗 2027.eacl.org/calls/srw/ ⏰ Pre-submission mentorship deadline: Nov 6, 2026 ⏰ Direct paper submission deadline: Dec 15, 2026 We accept both PhD thesis proposals and regular #nlproc research papers, looking forward to your submissions!

2027.eacl.org

Call for Student Research Workshop Papers

Official website for the 2027 Conference of the European Chapter of the Association for Computational Linguistics

075

Pasquale Minervini @neuralnoise.com · 02/08/2026

stochastic parrots are getting pretty lucky

050

Reposted by Pasquale Minervini

Ethan Mollick @emollick.bsky.social · 19/07/2026

I still believe that everyone is too fixated on the state of play in AI right now (which labs are ahead, how to manage costs, etc.) and not focused enough on the continued steepness of the capability curve for AI At higher capabilities (like the ones expected in the near term), a lot changes fast.

1014812

Pasquale Minervini @neuralnoise.com · 14/07/2026

Organisations -- sponsorship opportunities are available for #EACL2027, the flagship European conference in computational linguistics, taking place in Athens in March 2027! Support the NLP community and connect with researchers and practitioners: 2027.eacl.org

2027.eacl.org

The 20th Conference of the European Chapter of the Association for Computational LinguisticsAthens, GreeceMarch 9-14, 2027

8120

Reposted by Pasquale Minervini

Zeerak Talat زیرک طلعت (they/them) @zeerak.bsky.social · 08/07/2026

Thrilled to have been awarded the Association for Computational Linguistics 2016 test of time award for my first ever paper, written with/under the guidance of @dirkhovy.bsky.social A couple of cute things about the paper/its genesis/its outcomes

Photo of a test of time award for the paper "Hateful Symbols or Hateful People? Predictive Features for Hate Speech Detection on Twitter"

1612418

Pasquale Minervini @neuralnoise.com · 04/07/2026

New blog post on our {ICML, ACL} 2026 papers, plus some new interesting results on open-ended learning and multi-modal retrieval! neuralnoise.com/2026/icml-ac...

7170

Reposted by Pasquale Minervini

Zeerak Talat زیرک طلعت (they/them) @zeerak.bsky.social · 18/06/2026

Call for papers for special issue on Ethics in NLP in the journal Computational Linguistics. Discussions around ethics in #NLP and #ComputationalLinguistics is often limited by the expectations for what should be published in *CL conferences. Link to call: www.aclweb.org/portal/conte...

1109

Reposted by Pasquale Minervini

Zeerak Talat زیرک طلعت (they/them) @zeerak.bsky.social · 12/06/2026

NLP reviews suck! But why is not clear– @aclrollingreview.bsky.social provides *a lot* of guidance for how to review, but in that, first principles get lost. So @adamlopez.bsky.social and I have written down some of our thoughts on first principles.

medium.com

The Missing First Principles of Reviewing for ACL

Zeerak Talat & Adam Lopez, University of Edinburgh

1236

Reposted by Pasquale Minervini

EACL 2027 @eaclmeeting.bsky.social · 04/05/2026

Attention #NLProc researchers, the EACL 2027 website is officially LIVE: 2027.eacl.org! 🎉 🇬🇷 Join us in Athens, Greece (Mar 9-13, 2027) at #EACL2027 📅 ARR submission deadline: Aug 6, 2026. Open to all areas of CL/NLP + related fields. Stay tuned for the detailed CfP soon!

Picture of the Acropolis in Athens, Greece

23922

Pasquale Minervini @neuralnoise.com · 27/03/2026

If you are interested in privacy-preserving clinical NLP, we are recruiting a postdoc at the University of Edinburgh! The work is on LLMs/VLMs, AI privacy, and real-world health data in secure research environments. Apply by April 6th, 2026! More details: elxw.fa.em3.oraclecloud.com/hcmUI/Candid...

2102

Pasquale Minervini @neuralnoise.com · 17/03/2026

My amazing colleagues Sid and Michael are looking for a postdoc! 👇

130

Reposted by Pasquale Minervini

an-exlab.bsky.social @an-exlab.bsky.social · 17/03/2026

We are advertising a postdoc position to work on #generative #models, #structure #induction, and MI #estimation with Michael Gutmann as part of @genaihub.bsky.social ! elxw.fa.em3.oraclecloud.com/hcmUI/Candid... Get in touch! (#ML #AI) 👉 homepages.inf.ed.ac.uk/snaraya3/ 👉 michaelgutmann.github.io

homepages.inf.ed.ac.uk

Siddharth - Home

Sid's page

043

Pasquale Minervini @neuralnoise.com · 29/11/2025

Chatted with the amazing @elissawelle.bsky.social from @theverge.com about @rohit-saxena.bsky.social’s “Lost in Time” work (arxiv.org/abs/2502.05092) and much more! You can find the full article here 👇

010

Pasquale Minervini @neuralnoise.com · 17/10/2025

Check out Yu Zhao's (@yuzhaouoe.bsky.social) latest work, “Learning GUI Grounding with Spatial Reasoning from Visual Feedback” (www.arxiv.org/abs/2509.21552), done during his internship at MSR (@msftresearch.bsky.social)! New SOTA 🏆 results on ScreenSpot-v2 (+5.7%) and ScreenSpot-Pro (+110.8%)!

131

Reposted by Pasquale Minervini

mr. TIM @timkellogg.me · 24/08/2025

trend: non-NVIDIA training DeepSeek V3.1 was trained on Huawei Ascend NPUs this one is a South Korean lab training on AMD

2397

Pasquale Minervini @neuralnoise.com · 23/08/2025

I really needed a Deep Research MCP server to use with Claude Code and other tools — here it is: github.com/pminervini/d...

060

Reposted by Pasquale Minervini

Mark Riedl @markriedl.bsky.social · 05/08/2025

@togelius.bsky.social has thoughts on Genie 3 and games togelius.blogspot.com/2025/08/geni... Fairly close to my own, though I didn't get the preview the tech. Walking around a generated image-to-image world is not the same as playing a game. There are no game objectives.

togelius.blogspot.com

Genie 3 and the future of neural game engines

Google DeepMind just announced Genie 3 , their new promptable world model, which is another term for neural game engine. This is a big neura...

4153

Reposted by Pasquale Minervini

mr. TIM @timkellogg.me · 03/08/2025

quick diagram of Bluesky’s architecture and why it’s nicer here

diagram from Anthropic paper with an icon & label that says “subtract evil vector”

4725

Reposted by Pasquale Minervini

Scott McGrath @smcgrath.phd · 23/07/2025

Anthropic research identifies “inverse scaling in test-time compute,” where longer reasoning degrades AI performance. On certain tasks, models become more distracted by irrelevant data or overfit to spurious correlations. #MLSky

venturebeat.com

Anthropic researchers discover the weird AI problem: Why thinking longer makes models dumber

Anthropic research reveals AI models perform worse with extended reasoning time, challenging industry assumptions about test-time compute scaling in enterprise deployments.

191

Pasquale Minervini @neuralnoise.com · 31/07/2025

Supermassive congrats to Giwon Hong (@giwonhong.bsky.social) for the amazing feat! 🙂

141

Pasquale Minervini @neuralnoise.com · 26/07/2025

The amazing folks at EdinburghNLP will be presenting a few papers at ACL 2025 (@aclmeeting.bsky.social); if you're in Vienna, touch base with them!

0120

Reposted by Pasquale Minervini

Emile van Krieken @emilevankrieken.com · 24/07/2025

Hm, hard disagree here. I really fail to see how this is misconduct akin to bribery, it's just a defense mechanism against bad reviewing practices. @neuralnoise.com

152

Reposted by Pasquale Minervini

Sohee Yang @soheeyang.bsky.social · 13/06/2025

🚨 New Paper 🚨 How effectively do reasoning models reevaluate their thought? We find that: - Models excel at identifying unhelpful thoughts but struggle to recover from them - Smaller models can be more robust - Self-reevaluation ability is far from true meta-cognitive awareness 1/N 🧵

1123

Reposted by Pasquale Minervini

mr. TIM @timkellogg.me · 22/07/2025

Inverse scaling of reasoning models a research collab demonstrated that there are certain types of tasks where all top reasoning models do WORSE the longer they think things like getting distracted by irrelevant info, spurious correlations, etc. www.arxiv.org/abs/2507.14417

Three panels at the top describe task types with example prompts:
1. Simple Counting Tasks with Distractors (Misleading Math & Python):
• Prompts mention an apple and an orange, with added irrelevant or confusing information (e.g., probabilistic riddle, Python code) before asking the straightforward question: “Calculate how many fruits you have.”
2. Regression Tasks with Spurious Features (Grades Regression):
• Given XML-style records about a student, the model must predict grades from features like sleep hours, social hours, and stress level. The challenge lies in identifying relevant vs. spurious attributes.
3. Deduction Tasks with Constraint Tracking (Zebra Puzzles):
• Complex logical reasoning puzzle with multiple interrelated clues. Example: “What position is the person who likes salmon at?” Constraints involve foods, names, and relationships like “to the left of.”

Bottom row contains 3 line plots comparing model performance across tasks:
• Misleading Math (Left Plot):
• Accuracy drops sharply for some models as reasoning tokens increase. Claude Sonnet 4 maintains high performance. o3 and DeepSeek R1 hold relatively stable accuracy; Qwen3 32B and QwQ 32B drop more.
• Grades Regression (Middle Plot):
• Shows negative RMSE (higher is better). Claude models remain strong across token counts; o3 also performs well. Qwen3 and QwQ struggle, with DeepSeek R1 performing modestly.
• Zebra Puzzles (Right Plot):
• Accuracy vs. average reasoning tokens. o3 and Claude Sonnet 4 maintain highest performance. Other models (e.g., DeepSeek R1, Qwen3 32B, QwQ 32B) show performance degradation or plateaus. Error bars reflect variability.

Each plot uses colored lines with markers to indicate different model names.

2212

Reposted by Pasquale Minervini

Naomi Saphra @nsaphra.bsky.social · 12/06/2025

Reasoning is about variable binding. It’s not about information retrieval. If a model cannot do variable binding, it is not good at grounded reasoning, and there’s evidence accruing that large scale can make LLMs worse at in-context grounded reasoning. 🧵

4549

Pasquale Minervini @neuralnoise.com · 22/07/2025

Sometimes, too much reasoning can hurt model performance! New research by Anthropic (@anthropic.com), by Aryo Pradipta Gema (@aryopg.bsky.social) et al.: huggingface.co/papers/2507....

huggingface.co

Paper page - Inverse Scaling in Test-Time Compute

Join the discussion on this paper page

050

Pasquale Minervini @neuralnoise.com · 21/07/2025

“LLMs can’t reason” 😅

050

Reposted by Pasquale Minervini

Steven Strogatz @stevenstrogatz.com · 04/07/2025

My "Math, Revealed" series is freely available to anyone -- no paywall! -- in the thread below.

613653

Reposted by Pasquale Minervini

Cosimo Gregucci @cgregucci.bsky.social · 10/07/2025

Spotlight poster coming soon at #ICML2025 @icmlconf.bsky.social! 📌East Exhibition Hall A-B E-1806 🗓️Wed 16 Jul 4:30 p.m. PDT — 7 p.m. PDT 📜 arxiv.org/pdf/2410.12537 Let’s chat! I’m always up for conversations about knowledge graphs, reasoning, neuro-symbolic AI, and benchmarking.

1102

Reposted by Pasquale Minervini

Melanie Mitchell @melaniemitchell.bsky.social · 05/07/2025

This essay by Nisheeth Vishnoi is a thoughtful meditation on the nature of science and a rebuttal to the notion that AI systems are going replace human scientists anytime soon. Worth reading. nisheethvishnoi.substack.com/p/what-count...

nisheethvishnoi.substack.com

What Counts as Discovery?

Rethinking AI’s Place in Science

47312

Pasquale Minervini @neuralnoise.com · 05/07/2025

"in 2025 we will have flying cars" 😂😂😂

839991

Reposted by Pasquale Minervini

Bálint Gyevnár @gbalint.bsky.social · 30/05/2025

Preprint alert 🎉 Introducing the Agentic eXplanations via Interrogative Simulations (AXIS) algo. AXIS integrates multi-agent simulators with LLMs by having the LLMs interrogate the simulator with counterfactual queries over multiple rounds for explaining agent behaviour. arxiv.org/pdf/2505.17801

Flowchart of the AXIS algorithm with 5 parts. The top-left has the memory, the centre-left has the user query, the centre-bottom has the final explanation, the centre has the LLM, and the right has the multi-agent simulator.

081

Reposted by Pasquale Minervini

Bálint Gyevnár @gbalint.bsky.social · 17/04/2025

'AI Safety for Everyone' is out now in @natmachintell.nature.com! Through an analysis of 383 papers, we find a rich landscape of methods that cover a much larger domain than mainstream notions of AI safety. Our takeaway: Epistemic inclusivity is important, the knowledge is there, we only need use it

1133

Reposted by Pasquale Minervini

eleutherai.bsky.social @eleutherai.bsky.social · 06/06/2025

Can you train a performant language model using only openly licensed text? We are thrilled to announce the Common Pile v0.1, an 8TB dataset of openly licensed and public domain text. We train 7B models for 1T and 2T tokens and match the performance similar models like LLaMA 1 & 2

214660

Pasquale Minervini @neuralnoise.com · 04/06/2025

COLM (@colmweb.org‬) reviewers, please follow up on author responses if you need to! Most of the papers in my area chair batch didn't receive reviewer follow-ups, and it's dire

062

Pasquale Minervini @neuralnoise.com · 25/05/2025

claude-code is pretty good at updating personal websites! it has browser use, so it can e.g. scrape your latest papers from arxiv and dblp and use that to update your website's publication list

250

Pasquale Minervini @neuralnoise.com · 25/05/2025

“You must never be fearful about what you are doing when it is right.” -- Rosa Parks

050

Reposted by Pasquale Minervini

antonio vergari ⚔️ short-circuiting @nolovedeeplearning.bsky.social · 24/05/2025

🗓️ Deadline extended: 💥2nd June 2025!💥 We are looking forward to your works on: 🔌 #circuits and #tensor #networks 🕸️ ⏳ normalizing #flows 💨 ⚖️ scaling #NeSy #AI 🦕 🚅 fast and #reliable inference 🔍 ...& more! please share 🙏

01912

Reposted by Pasquale Minervini

Emile van Krieken @emilevankrieken.com · 22/05/2025

This is a new experience: people using AI to overhype your paper 🫢 @neuralnoise.com

191

Reposted by Pasquale Minervini

Emile van Krieken @emilevankrieken.com · 21/05/2025

We propose Neurosymbolic Diffusion Models! We find diffusion is especially compelling for neurosymbolic approaches, combining powerful multimodal understanding with symbolic reasoning 🚀 Read more 👇

49327

Reposted by Pasquale Minervini

Ellen Rhudy @ellenrhudy.bsky.social · 20/05/2025

Just realized that the AI slop also made it into the @inquirer.com. So glad I saved this list of novels that do not exist!!!

A summer reading list from the Philadelphia Inquirer, which consists almost entirely of nonexistent novels

2176

Reposted by Pasquale Minervini

Emile van Krieken @emilevankrieken.com · 17/05/2025

Congrats! Looks like time is a big failure case for these models (cc @neuralnoise.com @aryopg.bsky.social @rohit-saxena.bsky.social ) bsky.app/profile/emil...

132

Reposted by Pasquale Minervini

Aryo Pradipta Gema @aryopg.bsky.social · 02/05/2025

MMLU-Redux just touched down at #NAACL2025! 🎉 Wish I could be there for our "Are We Done with MMLU?" poster today (9:00-10:30am in Hall 3, Poster Session 7), but visa drama said nope 😅 If anyone's swinging by, give our research some love! Hit me up if you check it out! 👋

01711

Reposted by Pasquale Minervini

Ai2 @ai2.bsky.social · 01/05/2025

We're excited to round out the OLMo 2 family with its smallest member, OLMo 2 1B, surpassing peer models like Gemma 3 1B or Llama 3.2 1B. The 1B model should enable rapid iteration for researchers, more local development, and a more complete picture of how our recipe scales.

A bar graph comparing average performance (10 Tasks) across OLMo 2 1B, SmolLM2 1.7B, Gemma 3 1B, Llama 3.2 1B, and Qwen 2.5 1.5B. The highest performance is 42.7, achieved by OLMo 2 1B.

14210

Reposted by Pasquale Minervini

Yuchen Zhu @zhuyuchen.bsky.social · 01/05/2025

New work! 💪🏻💥🤯 When Can Proxies Improve the Sample Complexity of Preference Learning? Our paper is accepted at @icmlconf.bsky.social 2025. Fantastic joint work with @spectral.space, Zhengyan Shi, @meng-yue-yang.bsky.social, @neuralnoise.com, Matt Kusner, @alexdamour.bsky.social. 1/n

184

Pasquale Minervini @neuralnoise.com · 30/04/2025

Extremely proud of the squad 💕

030

Reposted by Pasquale Minervini

Manuel Dileo @manueldileo.bsky.social · 25/04/2025

Happy to share that our work won the best paper award at ESANN 2025! 🚀 Huge thanks to my amazing co-authors✨ Read our work here: doi.org/10.14428/esa...

031