If there is a time to apply the Robustness Principle, it's for earthquake alerts.... this is really sad.
October 17, 2026 at 7:31 AM
If there is a time to apply the Robustness Principle, it's for earthquake alerts.... this is really sad.
One week left to apply for ECR places to our workshop on developmental robustness!
Applications close on 23 October for funded ECR places at our Workshop on Bridging Scales and Disciplines to Understand Robust Development, 18-21 April 2027 organised by Mauricio Rocha Martins @rocha-lab.bsky.social and Giulia Paci @giuliapaci.bsky.social
Apply at biologists.com/workshops/ap...
Apply at biologists.com/workshops/ap...
October 16, 2026 at 9:32 AM
One week left to apply for ECR places to our workshop on developmental robustness!
Nate Hagen's conversation on society, resilience, robustness, performance, efficiency with French biologist #OlivierHamant is incredibly delightful, even joyful. Our #Ukraine documentary on robustness and climate in a time of conflict benefits! @natehagens.bsky.social
www.youtube.com/watch?v=Kzzr...
www.youtube.com/watch?v=Kzzr...
Why Too Much Efficiency Makes Us Fragile with Olivier Hamant | TGS 230
YouTube video by Nate Hagens
www.youtube.com
October 15, 2026 at 4:01 PM
Nate Hagen's conversation on society, resilience, robustness, performance, efficiency with French biologist #OlivierHamant is incredibly delightful, even joyful. Our #Ukraine documentary on robustness and climate in a time of conflict benefits! @natehagens.bsky.social
www.youtube.com/watch?v=Kzzr...
www.youtube.com/watch?v=Kzzr...
Biological brains "accept higher metabolic cost to gain both robustness and capability" www.biorxiv.org/content/10.6... Brains are expensive but they're good at what they do.
October 13, 2026 at 11:11 AM
Biological brains "accept higher metabolic cost to gain both robustness and capability" www.biorxiv.org/content/10.6... Brains are expensive but they're good at what they do.
this is brilliant from Zack Polanski. exactly the robustness we need. right to point out the injustice of locking up pensioners for speaking out against genocide
October 15, 2026 at 9:16 AM
this is brilliant from Zack Polanski. exactly the robustness we need. right to point out the injustice of locking up pensioners for speaking out against genocide
Biomimetic Molecular-Bridge Interfacial Engineering Simultaneously Enhances Sensitivity and Robustness in e-skin www.chinesechemsoc.org/doi/10.31635...
#chemistry #openaccess #science
#chemistry #openaccess #science
October 15, 2026 at 7:52 PM
Biomimetic Molecular-Bridge Interfacial Engineering Simultaneously Enhances Sensitivity and Robustness in e-skin www.chinesechemsoc.org/doi/10.31635...
#chemistry #openaccess #science
#chemistry #openaccess #science
Now in eLife: The effect of age on the neuronal power spectrum, with rigorous checks for reliability and robustness. We run control analyses and establish a framework for future study planning with effect sizes. Not all effects are well powered in the same sample...
doi.org/10.7554/eLif...
doi.org/10.7554/eLif...
October 14, 2026 at 8:57 AM
Now in eLife: The effect of age on the neuronal power spectrum, with rigorous checks for reliability and robustness. We run control analyses and establish a framework for future study planning with effect sizes. Not all effects are well powered in the same sample...
doi.org/10.7554/eLif...
doi.org/10.7554/eLif...
Fun with geo-located respondents.
The lack of detail about the models and demographics--and the absence of robustness tests--makes me skeptical.
Up next, 83 page Economist articles that are 5 years out of date
The lack of detail about the models and demographics--and the absence of robustness tests--makes me skeptical.
Up next, 83 page Economist articles that are 5 years out of date
Voters living near data centers are moving further toward Democrats
economist.com/graphic-deta...
economist.com/graphic-deta...
October 15, 2026 at 7:26 PM
Fun with geo-located respondents.
The lack of detail about the models and demographics--and the absence of robustness tests--makes me skeptical.
Up next, 83 page Economist articles that are 5 years out of date
The lack of detail about the models and demographics--and the absence of robustness tests--makes me skeptical.
Up next, 83 page Economist articles that are 5 years out of date
Call the robustness
October 16, 2026 at 2:00 AM
Call the robustness
New #TMLR-Paper-with-Video:
Do Object Channels Improve Robustness in Deep Reinforcement Learning?
Jannis Blüml, Cedric Derstroff, Bjarne Gregori et al.
https://tmlr.infinite-conf.org/paper_pages/7BFbso4B3R
#pixels #pixel #reinforcement
Do Object Channels Improve Robustness in Deep Reinforcement Learning?
Jannis Blüml, Cedric Derstroff, Bjarne Gregori et al.
https://tmlr.infinite-conf.org/paper_pages/7BFbso4B3R
#pixels #pixel #reinforcement
October 14, 2026 at 4:31 PM
New #TMLR-Paper-with-Video:
Do Object Channels Improve Robustness in Deep Reinforcement Learning?
Jannis Blüml, Cedric Derstroff, Bjarne Gregori et al.
https://tmlr.infinite-conf.org/paper_pages/7BFbso4B3R
#pixels #pixel #reinforcement
Do Object Channels Improve Robustness in Deep Reinforcement Learning?
Jannis Blüml, Cedric Derstroff, Bjarne Gregori et al.
https://tmlr.infinite-conf.org/paper_pages/7BFbso4B3R
#pixels #pixel #reinforcement
"These systems decide which experiments to report using rewards that directly observe performance on the held-out data, preferentially elevating results that perform well on final evaluation targets, even when earlier evidence of robustness or generalization is weak."
October 15, 2026 at 5:51 PM
"These systems decide which experiments to report using rewards that directly observe performance on the held-out data, preferentially elevating results that perform well on final evaluation targets, even when earlier evidence of robustness or generalization is weak."
SO MANY commenters here confuse the need to ensure that rulings are robust to appeal with tepidity.
And that robustness — that rigor — is the very robustness and rigor that protects the innocent and is an essential prerequisite for fair rulings in other cases.
And that robustness — that rigor — is the very robustness and rigor that protects the innocent and is an essential prerequisite for fair rulings in other cases.
October 14, 2026 at 7:22 PM
SO MANY commenters here confuse the need to ensure that rulings are robust to appeal with tepidity.
And that robustness — that rigor — is the very robustness and rigor that protects the innocent and is an essential prerequisite for fair rulings in other cases.
And that robustness — that rigor — is the very robustness and rigor that protects the innocent and is an essential prerequisite for fair rulings in other cases.
The new Technical Designer will “ensure the technical excellence of all #Warhammer40k rules” through “mathematical and logical robustness”. What would you do in their shoes? www.wargamer.com/warhammer-40...
Games Workshop seeks math-minded designer to make Warhammer 40k "fair, balanced, and enjoyable"
The new Technical Designer will "ensure the technical excellence of all Warhammer 40,000 rules" through "mathematical and logical robustness".
www.wargamer.com
October 13, 2026 at 1:22 PM
The new Technical Designer will “ensure the technical excellence of all #Warhammer40k rules” through “mathematical and logical robustness”. What would you do in their shoes? www.wargamer.com/warhammer-40...
Noise, bias and data limitations: Virtual benchmarking of species distribution models' robustness for prediction and inference www.sciencedirect.com/science/arti...
Noise, bias and data limitations: Virtual benchmarking of species distribution models' robustness for prediction and inference
Species distribution models (SDMs) are widely used to describe and predict species occurrences from environmental data, but calibration datasets are t…
www.sciencedirect.com
October 14, 2026 at 12:52 PM
Noise, bias and data limitations: Virtual benchmarking of species distribution models' robustness for prediction and inference www.sciencedirect.com/science/arti...
No, the robustness of the record also matters, and it matters a *lot*. The more robust, the smaller the subset of judges who are willing to look the other way.
October 12, 2026 at 8:12 PM
No, the robustness of the record also matters, and it matters a *lot*. The more robust, the smaller the subset of judges who are willing to look the other way.
#AI can make #ocean predictions more reliable by embedding #physical knowledge, though uncertainty and robustness still need further work.👇
@olarjournal.bsky.social
spj.science.org/doi/epdf/10....
@olarjournal.bsky.social
spj.science.org/doi/epdf/10....
October 16, 2026 at 9:38 AM
#AI can make #ocean predictions more reliable by embedding #physical knowledge, though uncertainty and robustness still need further work.👇
@olarjournal.bsky.social
spj.science.org/doi/epdf/10....
@olarjournal.bsky.social
spj.science.org/doi/epdf/10....
One day God willing we will have a normie scotus who will overturn NetChoice v. Paxton and a Congress who will pass a Digital First Amendment Robustness Act.
October 12, 2026 at 4:30 PM
One day God willing we will have a normie scotus who will overturn NetChoice v. Paxton and a Congress who will pass a Digital First Amendment Robustness Act.
AgentPerturb: Evaluating LLM Agent Robustness under Controlled Interaction Perturbations
Yuanjian Xu et al.
#arXiv #cs.AI
Yuanjian Xu et al.
#arXiv #cs.AI
AgentPerturb: Evaluating LLM Agent Robustness under Controlled Interaction Perturbations
Real-world tasks are often complex and noisy, posing substantial challenges for LLM agents. Yet existing evaluations lack a systematic framework for jointly characterizing potential perturbation targets and types. We use this two-dimensional view to conduct a unified study that quantifies the robus…
arxiv.org
October 17, 2026 at 7:43 AM
Safe autonomy demands tested contingencies—fallback architectures, mission abort logic, and recovery procedures. Without them, lost connectivity becomes a command-and-control failure, proving robustness needs more than just links.
October 17, 2026 at 8:45 AM
Safe autonomy demands tested contingencies—fallback architectures, mission abort logic, and recovery procedures. Without them, lost connectivity becomes a command-and-control failure, proving robustness needs more than just links.
Measuring AI Reliability Through Robustness, Safety, and Governance
AI models can ace every benchmark and still cause harm. Here is a practical, layered way to measure safety, robustness and governance, not just accuracy.
#hackernews #news
AI models can ace every benchmark and still cause harm. Here is a practical, layered way to measure safety, robustness and governance, not just accuracy.
#hackernews #news
Measuring AI Reliability Through Robustness, Safety, and Governance
AI models can ace every benchmark and still cause harm. Here is a practical, layered way to measure safety, robustness and governance, not just accuracy.
hackernoon.com
October 17, 2026 at 6:59 AM
Measuring AI Reliability Through Robustness, Safety, and Governance
AI models can ace every benchmark and still cause harm. Here is a practical, layered way to measure safety, robustness and governance, not just accuracy.
#hackernews #news
AI models can ace every benchmark and still cause harm. Here is a practical, layered way to measure safety, robustness and governance, not just accuracy.
#hackernews #news
AgentPerturb: Evaluating LLM Agent Robustness under Controlled Interaction Perturbations
Yuanjian Xu et al.
#arXiv #cs.AI
Yuanjian Xu et al.
#arXiv #cs.AI
AgentPerturb: Evaluating LLM Agent Robustness under Controlled Interaction Perturbations
Real-world tasks are often complex and noisy, posing substantial challenges for LLM agents. Yet existing evaluations lack a systematic framework for jointly characterizing potential perturbation targets and types. We use this two-dimensional view to conduct a unified study that quantifies the robus…
arxiv.org
October 17, 2026 at 5:43 AM