Mastodon 🐘

HackerNoonPEAR: a novel loss term enhances interpretability and confidence in deep learning models for tabular data, hence boosting consensus in AI explainers. <a href="https://hackernoon.com/the-geeks-guide-to-ml-experimentation" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/the-geeks-guide-to-ml-experimentation</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

HackerNoonDiscover how PEAR increases explainer consensus, paving the way for deep learning models that are fair, interpretable, and prepared for the future. <a href="https://hackernoon.com/can-pear-make-deep-learning-easier-to-trust" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/can-pear-make-deep-learning-easier-to-trust</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

HackerNoonLearn how PEAR exhibits optimal consensus when both loss terms are balanced, improves linearity, and preserves useful, non-trivial explanations. <a href="https://hackernoon.com/consensus-loss-proves-ai-can-be-both-accurate-and-transparent" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/consensus-loss-proves-ai-can-be-both-accurate-and-transparent</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

HackerNoonLearn how PEAR offers interpretability improvements with negligible accuracy tradeoffs by improving explainer agreement across measures and invisible explainers <a href="https://hackernoon.com/the-trade-off-between-accuracy-and-agreement-in-ai-models" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/the-trade-off-between-accuracy-and-agreement-in-ai-models</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

HackerNoonDiscover how to reduce disagreement across feature attributions, PEAR uses correlation-based loss to train neural networks for accuracy and explainer consensus. <a href="https://hackernoon.com/notes-on-training-neural-networks-for-consensus" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/notes-on-training-neural-networks-for-consensus</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

HackerNoonLearn how PEAR balances accuracy and interpretability by increasing post hoc explanation consensus and decreasing feature attribution disagreement. <a href="https://hackernoon.com/new-ai-study-tackles-the-transparency-problem-in-black-box-models" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/new-ai-study-tackles-the-transparency-problem-in-black-box-models</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

FIZ ISE Research GroupDo you already know our FIZ ISE Zenodo community? Find the latest publications of our research group including presentations and posters here: <a href="https://zenodo.org/communities/fizise/records" rel="nofollow noopener" translate="no" target="_blank">https://zenodo.org/communities/fizise/records</a><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://sigmoid.social/tags/knowledgegraphs" class="mention hashtag" rel="nofollow noopener" target="_blank">#knowledgegraphs</a> <a href="https://sigmoid.social/tags/semanticweb" class="mention hashtag" rel="nofollow noopener" target="_blank">#semanticweb</a> <a href="https://sigmoid.social/tags/explainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableAI</a> <a href="https://nfdi.social/@nfdi4culture" class="u-url mention" rel="nofollow noopener" target="_blank">@nfdi4culture</a> <a href="https://nfdi.social/@NFDI4DS" class="u-url mention" rel="nofollow noopener" target="_blank">@NFDI4DS</a> <a href="https://nfdi.social/@NFDI4Memory" class="u-url mention" rel="nofollow noopener" target="_blank">@NFDI4Memory</a> <a href="https://fedihum.org/@tabea" class="u-url mention" rel="nofollow noopener" target="_blank">@tabea</a> <a href="https://fedihum.org/@sashabruns" class="u-url mention" rel="nofollow noopener" target="_blank">@sashabruns</a> <a href="https://sigmoid.social/@AnnaJacyszyn" class="u-url mention" rel="nofollow noopener" target="_blank">@AnnaJacyszyn</a> <a href="https://sigmoid.social/tags/DiTraRe" class="mention hashtag" rel="nofollow noopener" target="_blank">#DiTraRe</a> <a href="https://sigmoid.social/@enorouzi" class="u-url mention" rel="nofollow noopener" target="_blank">@enorouzi</a> <a href="https://blog.epoz.org/" class="u-url mention" rel="nofollow noopener" target="_blank">@epoz</a> <a href="https://sigmoid.social/@GenAsefa" class="u-url mention" rel="nofollow noopener" target="_blank">@GenAsefa</a> <a href="https://sigmoid.social/@joerg" class="u-url mention" rel="nofollow noopener" target="_blank">@joerg</a> <a href="https://sigmoid.social/tags/pmd" class="mention hashtag" rel="nofollow noopener" target="_blank">#pmd</a> <a href="https://sigmoid.social/tags/materialsscience" class="mention hashtag" rel="nofollow noopener" target="_blank">#materialsscience</a> <a href="https://sigmoid.social/tags/MSE" class="mention hashtag" rel="nofollow noopener" target="_blank">#MSE</a> <a href="https://sigmoid.social/tags/NFDIMatWerk" class="mention hashtag" rel="nofollow noopener" target="_blank">#NFDIMatWerk</a> <a href="https://sigmoid.social/tags/Wiedergutmachung" class="mention hashtag" rel="nofollow noopener" target="_blank">#Wiedergutmachung</a> <a href="https://sigmoid.social/@MahsaVafaie" class="u-url mention" rel="nofollow noopener" target="_blank">@MahsaVafaie</a> <a href="https://fosstodon.org/@heikef" class="u-url mention" rel="nofollow noopener" target="_blank">@heikef</a> <a href="https://wisskomm.social/@fiz_karlsruhe" class="u-url mention" rel="nofollow noopener" target="_blank">@fiz_karlsruhe</a>

DiTraReUnlock the next wave of AI-driven chemical discoveries with the Chemotion Knowledge Graph, a new open-source pipeline making lab data truly FAIR and intelligent. <a href="https://sigmoid.social/@enorouzi" class="u-url mention" rel="nofollow noopener" target="_blank">@enorouzi</a> Nicole Jung, <a href="https://sigmoid.social/@AnnaJacyszyn" class="u-url mention" rel="nofollow noopener" target="_blank">@AnnaJacyszyn</a> <a href="https://sigmoid.social/@joerg" class="u-url mention" rel="nofollow noopener" target="_blank">@joerg</a> and <a href="https://sigmoid.social/@lysander07" class="u-url mention" rel="nofollow noopener" target="_blank">@lysander07</a> AI4DiTraRe: Building the BFO-Compliant Chemotion Knowledge Graph to be published at Sci-K 2025 workshop at <a href="https://mastodon.social/@iswc_conf" class="u-url mention" rel="nofollow noopener" target="_blank">@iswc_conf</a> ' <a href="https://zenodo.org/records/17046436" rel="nofollow noopener" translate="no" target="_blank">https://zenodo.org/records/17046436</a><a href="https://social.kit.edu/tags/chemistry" class="mention hashtag" rel="nofollow noopener" target="_blank">#chemistry</a> <a href="https://nfdi.social/@NFDI4Chem" class="u-url mention" rel="nofollow noopener" target="_blank">@NFDI4Chem</a> <a href="https://social.kit.edu/tags/ditrare" class="mention hashtag" rel="nofollow noopener" target="_blank">#ditrare</a> <a href="https://social.kit.edu/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://social.kit.edu/tags/knowledgegraphs" class="mention hashtag" rel="nofollow noopener" target="_blank">#knowledgegraphs</a> <a href="https://social.kit.edu/tags/semanticweb" class="mention hashtag" rel="nofollow noopener" target="_blank">#semanticweb</a> <a href="https://social.kit.edu/tags/explainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableAI</a> <a href="https://social.kit.edu/tags/chemotion" class="mention hashtag" rel="nofollow noopener" target="_blank">#chemotion</a>

Harald SackThe SEMANTiCS 2025 agenda is packed: keynotes, workshops, paper sessions, posters, and networking opportunities every day. You can still register for on site or online participation 👉 <a href="https://2025-eu.semantics.cc" rel="nofollow noopener" translate="no" target="_blank">https://2025-eu.semantics.cc</a><a href="https://sigmoid.social/tags/SemanticsConf" class="mention hashtag" rel="nofollow noopener" target="_blank">#SemanticsConf</a> <a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://sigmoid.social/tags/SemanticWeb" class="mention hashtag" rel="nofollow noopener" target="_blank">#SemanticWeb</a> <a href="https://sigmoid.social/tags/knowledgegraphs" class="mention hashtag" rel="nofollow noopener" target="_blank">#knowledgegraphs</a> <a href="https://sigmoid.social/tags/explainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableAI</a> <a href="https://sigmoid.social/tags/reliableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#reliableAI</a> <a href="https://sigmoid.social/tags/llms" class="mention hashtag" rel="nofollow noopener" target="_blank">#llms</a>

Sanjay MohindrooExplainable AI is reshaping trust, compliance, and leadership in digital transformation. Ready to shift from black-box to glass-box models? <a href="https://social.vivaldi.net/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> <a href="https://social.vivaldi.net/tags/DigitalLeadership" class="mention hashtag" rel="nofollow noopener" target="_blank">#DigitalLeadership</a> <a href="https://social.vivaldi.net/tags/RiskManagement" class="mention hashtag" rel="nofollow noopener" target="_blank">#RiskManagement</a> <a href="https://social.vivaldi.net/tags/TechStrategy" class="mention hashtag" rel="nofollow noopener" target="_blank">#TechStrategy</a> <a href="https://social.vivaldi.net/tags/CIOPriorities" class="mention hashtag" rel="nofollow noopener" target="_blank">#CIOPriorities</a> <a href="https://social.vivaldi.net/tags/AICompliance" class="mention hashtag" rel="nofollow noopener" target="_blank">#AICompliance</a> <a href="https://social.vivaldi.net/tags/XAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#XAI</a> <a href="https://social.vivaldi.net/tags/AITrust" class="mention hashtag" rel="nofollow noopener" target="_blank">#AITrust</a> <a href="https://social.vivaldi.net/tags/DataGovernance" class="mention hashtag" rel="nofollow noopener" target="_blank">#DataGovernance</a> <a href="https://social.vivaldi.net/tags/ITTransformation" class="mention hashtag" rel="nofollow noopener" target="_blank">#ITTransformation</a> <a href="https://social.vivaldi.net/tags/EmergingTech" class="mention hashtag" rel="nofollow noopener" target="_blank">#EmergingTech</a> <a href="https://social.vivaldi.net/tags/SanjayKMohindroo" class="mention hashtag" rel="nofollow noopener" target="_blank">#SanjayKMohindroo</a> <a href="https://medium.com/@sanjay.mohindroo66/the-rise-of-explainable-ai-xai-and-its-role-in-risk-management-792b5df68902" rel="nofollow noopener" translate="no" target="_blank">https://medium.com/@sanjay.mohindroo66/the-rise-of-explainable-ai-xai-and-its-role-in-risk-management-792b5df68902</a>

HackerNoonThis deep-dive analysis proves why SymTax works, showing its 'symbiotic' enricher and taxonomy fusion are essential for its state-of-the-art performance. <a href="https://hackernoon.com/a-quantitative-and-qualitative-analysis-of-the-symtax-citation-recommendation-model" rel="nofollow noopener" translate="no" target="_blank">https://hackernoon.com/a-quantitative-and-qualitative-analysis-of-the-symtax-citation-recommendation-model</a> <a href="https://mas.to/tags/explainableai" class="mention hashtag" rel="nofollow noopener" target="_blank">#explainableai</a>

Philo Sophies<a href="https://techhub.social/tags/Zoomposium" class="mention hashtag" rel="nofollow noopener" target="_blank">#Zoomposium</a> with Dr. <a href="https://techhub.social/tags/Patrick" class="mention hashtag" rel="nofollow noopener" target="_blank">#Patrick</a> <a href="https://techhub.social/tags/Krau%C3%9F" class="mention hashtag" rel="nofollow noopener" target="_blank">#Krauß</a>: “Instructions for building <a href="https://techhub.social/tags/artificialconsciousness" class="mention hashtag" rel="nofollow noopener" target="_blank">#artificialconsciousness</a>”Translating the various stages of <a href="https://techhub.social/tags/Damasio" class="mention hashtag" rel="nofollow noopener" target="_blank">#Damasio</a> <a href="https://techhub.social/tags/theoryofconsciousness" class="mention hashtag" rel="nofollow noopener" target="_blank">#theoryofconsciousness</a> 1:1 into concrete <a href="https://techhub.social/tags/circuitdiagrams" class="mention hashtag" rel="nofollow noopener" target="_blank">#circuitdiagrams</a> for <a href="https://techhub.social/tags/deeplearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#deeplearning</a>. To this end, strategies such as <a href="https://techhub.social/tags/feedforwardconnections" class="mention hashtag" rel="nofollow noopener" target="_blank">#feedforwardconnections</a> and <a href="https://techhub.social/tags/recurrentconnections" class="mention hashtag" rel="nofollow noopener" target="_blank">#recurrentconnections</a> in the form of <a href="https://techhub.social/tags/reinforcementlearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#reinforcementlearning</a> and <a href="https://techhub.social/tags/unsupervisedlearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#unsupervisedlearning</a> are used to simulate the <a href="https://techhub.social/tags/biologicalprocesses" class="mention hashtag" rel="nofollow noopener" target="_blank">#biologicalprocesses</a> of <a href="https://techhub.social/tags/neuralnetworks" class="mention hashtag" rel="nofollow noopener" target="_blank">#neuralnetworks</a>. More at: <a href="https://philosophies.de/index.php/2023/10/24/bauanleitung-kuenstliches-bewusstsein/" rel="nofollow noopener" translate="no" target="_blank">https://philosophies.de/index.php/2023/10/24/bauanleitung-kuenstliches-bewusstsein/</a>or: <a href="https://open.spotify.com/episode/0httDEYGoUGxJ6lHA03LEZ?si=BzmZn3EvQr6imSzUUKG8RQ" rel="nofollow noopener" translate="no" target="_blank">https://open.spotify.com/episode/0httDEYGoUGxJ6lHA03LEZ?si=BzmZn3EvQr6imSzUUKG8RQ</a><a href="https://techhub.social/tags/PatrickKrau%C3%9F" class="mention hashtag" rel="nofollow noopener" target="_blank">#PatrickKrauß</a> <a href="https://techhub.social/tags/PatrickKrauss" class="mention hashtag" rel="nofollow noopener" target="_blank">#PatrickKrauss</a> <a href="https://techhub.social/tags/Zoomposium" class="mention hashtag" rel="nofollow noopener" target="_blank">#Zoomposium</a> <a href="https://techhub.social/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#ArtificialIntelligence</a> <a href="https://techhub.social/tags/Consciousness" class="mention hashtag" rel="nofollow noopener" target="_blank">#Consciousness</a> <a href="https://techhub.social/tags/DeepLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#DeepLearning</a> <a href="https://techhub.social/tags/Neuroscience" class="mention hashtag" rel="nofollow noopener" target="_blank">#Neuroscience</a> <a href="https://techhub.social/tags/Damasio" class="mention hashtag" rel="nofollow noopener" target="_blank">#Damasio</a> <a href="https://techhub.social/tags/ArtificialConsciousness" class="mention hashtag" rel="nofollow noopener" target="_blank">#ArtificialConsciousness</a> <a href="https://techhub.social/tags/NeuroscienceInspiredAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#NeuroscienceInspiredAI</a> <a href="https://techhub.social/tags/CognitiveNeuroscience" class="mention hashtag" rel="nofollow noopener" target="_blank">#CognitiveNeuroscience</a> <a href="https://techhub.social/tags/Embodiment" class="mention hashtag" rel="nofollow noopener" target="_blank">#Embodiment</a> <a href="https://techhub.social/tags/TuringTest" class="mention hashtag" rel="nofollow noopener" target="_blank">#TuringTest</a> <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://techhub.social/tags/ArtificialConsciousness" class="mention hashtag" rel="nofollow noopener" target="_blank">#ArtificialConsciousness</a> <a href="https://techhub.social/tags/Ethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#Ethics</a> <a href="https://techhub.social/tags/AIrisks" class="mention hashtag" rel="nofollow noopener" target="_blank">#AIrisks</a> <a href="https://techhub.social/tags/MachineConsciousness" class="mention hashtag" rel="nofollow noopener" target="_blank">#MachineConsciousness</a> <a href="https://techhub.social/tags/ComputationalNeuroscience" class="mention hashtag" rel="nofollow noopener" target="_blank">#ComputationalNeuroscience</a> <a href="https://techhub.social/tags/ReinforcementLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#ReinforcementLearning</a> <a href="https://techhub.social/tags/UnsupervisedLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#UnsupervisedLearning</a> <a href="https://techhub.social/tags/CognitiveComputing" class="mention hashtag" rel="nofollow noopener" target="_blank">#CognitiveComputing</a> <a href="https://techhub.social/tags/ArtificialGeneralIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#ArtificialGeneralIntelligence</a> <a href="https://techhub.social/tags/PhilosophyOfMind" class="mention hashtag" rel="nofollow noopener" target="_blank">#PhilosophyOfMind</a> <a href="https://techhub.social/tags/BlackBoxProblem" class="mention hashtag" rel="nofollow noopener" target="_blank">#BlackBoxProblem</a> <a href="https://techhub.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> <a href="https://techhub.social/tags/Neuroinformatics" class="mention hashtag" rel="nofollow noopener" target="_blank">#Neuroinformatics</a> <a href="https://techhub.social/tags/DigitalMind" class="mention hashtag" rel="nofollow noopener" target="_blank">#DigitalMind</a> <a href="https://techhub.social/tags/FutureOfAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#FutureOfAI</a>

WebHeads UnitedAs AI tackles more complex problems, users need to know the 'why' behind the 'what'.An analytical tone is the voice of Explainable AI (XAI). It's objective, data-driven, and structured to clearly communicate its reasoning. This isn't just a tone choice; it's a foundation for building user trust and transparency in data-heavy applications.This guide covers the art and science of this essential AI persona.Read it here: <a href="https://webheadsunited.com/art-and-science-analytical-tone-in-ai-personas/" rel="nofollow noopener" translate="no" target="_blank">https://webheadsunited.com/art-and-science-analytical-tone-in-ai-personas/</a><a href="https://mastodon.social/tags/AIPersona" class="mention hashtag" rel="nofollow noopener" target="_blank">#AIPersona</a> <a href="https://mastodon.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> <a href="https://mastodon.social/tags/XAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#XAI</a> <a href="https://mastodon.social/tags/DataScience" class="mention hashtag" rel="nofollow noopener" target="_blank">#DataScience</a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a>

Doug OrtizThe rise of AI-powered database interaction is changing the dev game! 🤯 But, is the inherent lack of transparency in AI-generated SQL queries a ticking time bomb for security? What measures should developers prioritize to ensure the reliability and safety of these systems?Blog post: <a href="https://dougortiz.blogspot.com/2025/07/the-ai-database-interaction-paradox-why.html" rel="nofollow noopener" translate="no" target="_blank">https://dougortiz.blogspot.com/2025/07/the-ai-database-interaction-paradox-why.html</a><a href="https://mastodon.social/tags/Aidrivendatabaseinteraction" class="mention hashtag" rel="nofollow noopener" target="_blank">#Aidrivendatabaseinteraction</a> <a href="https://mastodon.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> <a href="https://mastodon.social/tags/AISecurity" class="mention hashtag" rel="nofollow noopener" target="_blank">#AISecurity</a> <a href="https://mastodon.social/tags/DeepTech" class="mention hashtag" rel="nofollow noopener" target="_blank">#DeepTech</a>

Anita Graser 🇪🇺🇺🇦🇬🇪Very excited that our most recent paper on <a href="https://fosstodon.org/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> & <a href="https://fosstodon.org/tags/ActiveLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#ActiveLearning</a> has been accepted to the <a href="https://fosstodon.org/tags/SENTIS" class="mention hashtag" rel="nofollow noopener" target="_blank">#SENTIS</a> workshop at <a href="https://fosstodon.org/tags/semanticsconf" class="mention hashtag" rel="nofollow noopener" target="_blank">#semanticsconf</a> by <a href="https://sigmoid.social/@semantics" class="u-url mention" rel="nofollow noopener" target="_blank">@semantics</a> <a href="https://semsys.ai.wu.ac.at/sentis2025/" rel="nofollow noopener" translate="no" target="_blank">https://semsys.ai.wu.ac.at/sentis2025/</a>

WinbuzzerMicrosoft and UW Develop AI That Spots Breast Cancer by Learning What's Normal<a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://mastodon.social/tags/Microsoft" class="mention hashtag" rel="nofollow noopener" target="_blank">#Microsoft</a> <a href="https://mastodon.social/tags/MedicalAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#MedicalAI</a> <a href="https://mastodon.social/tags/Healthcare" class="mention hashtag" rel="nofollow noopener" target="_blank">#Healthcare</a> <a href="https://mastodon.social/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#MachineLearning</a> <a href="https://mastodon.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a><a href="https://winbuzzer.com/2025/07/24/microsoft-and-uw-develop-ai-that-spots-breast-cancer-by-learning-whats-normal-xcxwbn" rel="nofollow noopener" translate="no" target="_blank">https://winbuzzer.com/2025/07/24/microsoft-and-uw-develop-ai-that-spots-breast-cancer-by-learning-whats-normal-xcxwbn</a>

BIG-5 ProjectHuge congratulations to Paula Feliu Criado for the successful defense of her Bachelor's thesis within the BIG-5 Project 🎓 Paula’s work developed a multitask AI pipeline to analyze how we express our connection to nature on social media. Going beyond simple classification, she used Explainable AI to understand how the model "sees" images, ensuring its insights are transparent and trustworthy.👉 Read more about her work on our blog: <a href="https://bit.ly/3Uotk4S" rel="nofollow noopener" translate="no" target="_blank">https://bit.ly/3Uotk4S</a><a href="https://mastodon.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> <a href="https://mastodon.social/tags/EnvironmentalScience" class="mention hashtag" rel="nofollow noopener" target="_blank">#EnvironmentalScience</a>

Aneesh SatheAI: Explainable EnoughThey look really juicy, she said. I was sitting in a small room with a faint chemical smell, doing one my first customer interviews. There is a sweet spot between going too deep and asserting a position. Good AI has to be just explainable enough to satisfy the user without overwhelming them with information. Luckily, I wasn’t new to the problem. <a href="https://aneeshsathe.com/wp-content/uploads/2025/07/image-from-rawpixel-id-3045306-jpeg.jpg" rel="nofollow noopener" target="_blank"></a>Nuthatcher atop Persimmons (ca. 1910) by Ohara Koson. Original from The Clark Art Institute. Digitally enhanced by rawpixel.Coming from a microscopy and bio background with a strong inclination towards image analysis I had picked up deep learning as a way to be lazy in lab. Why bother figuring out features of interest when you can have a computer do it for you, was my angle. The issue was that in 2015 no biologist would accept any kind of deep learning analysis and definitely not if you couldn’t explain the details. What the domain expert user doesn’t want: – How a convolutional neural network works. Confidence scores, loss, AUC, are all meaningless to a biologist and also to a doctor. What the domain expert desires:  – Help at the lowest level of detail that they care about.  – AI identifies features A, B, C, and that when you see A, B, & C it is likely to be disease X. Most users don’t care how a deep learning really works. So, if you start giving them details like the IoU score of the object detection bounding box or if it was YOLO or R-CNN that you used their eyes will glaze over and you will never get a customer. Draw a bounding box, heat map, or outline, with the predicted label and stop there. It’s also bad to go to the other extreme. If the AI just states the diagnosis for the whole image then the AI might be right, but the user does not get to participate in the process. Not to mention regulatory risk goes way up.This applies beyong images, consider LLMs. No one with any expertise likes a black box. Today, why do LLMs generate code instead of directly doing the thing that the programmer is asking them to do? It’s because the programmer wants to ensure that the code “works” and they have the expertise to figure out if and when it goes wrong. It’s the same reason that vibe coding is great for prototyping but not for production and why frequent readers can spot AI patterns, ahem,  easily.  So in a Betty Crocker cake mix kind of way, let the user add the egg. Building explainable-enough AI takes immense effort. It actually is easier to train AI to diagnose the whole image or to give details. Generating high-quality data at that just right level is very difficult and expensive. However, do it right and the effort pays off. The outcome is an AI-Human causal prediction machine. Where the causes, i.e. the median level features, inform the user and build confidence towards the final outcome. The deep learning part is still a black box but the user doesn’t mind because you aid their thinking. I’m excited by some new developments like <a href="https://rex-xai.readthedocs.io/en/stable/" rel="nofollow noopener" target="_blank">REX</a> which sort of retro-fit causality onto usual deep learning models. With improvements in performance user preferences for detail may change, but I suspect that need for AI to be explainable enough will remain. Perhaps we will even have custom labels like ‘juicy’.<a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai/" target="_blank">#AI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-adoption/" target="_blank">#AIAdoption</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-communication/" target="_blank">#AICommunication</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-explainability/" target="_blank">#AIExplainability</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-for-doctors/" target="_blank">#AIForDoctors</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-in-healthcare/" target="_blank">#AIInHealthcare</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-in-the-wild/" target="_blank">#AIInTheWild</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-product-design/" target="_blank">#AIProductDesign</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/ai-ux/" target="_blank">#AIUX</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/artificial-intelligence/" target="_blank">#artificialIntelligence</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/betty-crocker-thinking/" target="_blank">#BettyCrockerThinking</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/biomedical-ai/" target="_blank">#BiomedicalAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/business/" target="_blank">#Business</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/causal-ai/" target="_blank">#CausalAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/data-product-design/" target="_blank">#DataProductDesign</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/deep-learning/" target="_blank">#DeepLearning</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/explainable-ai/" target="_blank">#ExplainableAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/human-ai-interaction/" target="_blank">#HumanAIInteraction</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/image-analysis/" target="_blank">#ImageAnalysis</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/llms/" target="_blank">#LLMs</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/machine-learning-2/" target="_blank">#MachineLearning</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/startup-lessons/" target="_blank">#StartupLessons</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/statistics/" target="_blank">#statistics</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/tech-metaphors/" target="_blank">#TechMetaphors</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/tech-philosophy/" target="_blank">#techPhilosophy</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/trust-in-ai/" target="_blank">#TrustInAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/user-centered-ai/" target="_blank">#UserCenteredAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://aneeshsathe.com/tag/xai/" target="_blank">#XAI</a>

IJCAI ConferenceExcited to welcome Cynthia Rudin, Duke University, as an invited speaker at <a href="https://mastodon.social/tags/IJCAI2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#IJCAI2025</a> in Montreal! 🇨🇦 Known for her groundbreaking work in interpretable machine learning, Prof. Rudin will deliver her IJCAI-25 John McCarthy Award lecture. 🎥Why <a href="https://mastodon.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a>? <a href="https://youtu.be/9kzO5CKzFxQ" rel="nofollow noopener" translate="no" target="_blank">https://youtu.be/9kzO5CKzFxQ</a>

Vojtech CahlikThe paper, code, and data are available here: <a href="https://cahlik.net/reasoning-grounded-explanations-paper/" rel="nofollow noopener" translate="no" target="_blank">https://cahlik.net/reasoning-grounded-explanations-paper/</a><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://sigmoid.social/tags/genAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#genAI</a> <a href="https://sigmoid.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#LLM</a> <a href="https://sigmoid.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#LLMs</a> <a href="https://sigmoid.social/tags/ExplainableAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#ExplainableAI</a> <a href="https://sigmoid.social/tags/AIsafety" class="mention hashtag" rel="nofollow noopener" target="_blank">#AIsafety</a> <a href="https://sigmoid.social/tags/NLP" class="mention hashtag" rel="nofollow noopener" target="_blank">#NLP</a> <a href="https://sigmoid.social/tags/ML" class="mention hashtag" rel="nofollow noopener" target="_blank">#ML</a>

Recent searches

Search options

Administered by:

Server stats:

#explainableai