#robustai 搜尋結果

AI’s greatest strength—its ability to learn—can also be its biggest weakness. Adversarial ML exploits this, turning tiny tweaks into massive threats. From crashing cars to tricking diagnoses, the risks are real. Time to fight back. #AIsecurity #AdversarialML #RobustAI

7️⃣/🔟 🌍🤹 Real world = surprises. Tray bumped? ⚖️ Controller stabilizes NOW 🧠 Planner updates NEXT Grace under chaos. 🫡 #Safety #Realtime #RobustAI

The Pentagon’s Joint AI Center publishes annual AI ethics metrics—documenting bias tests, robustness checks, and explainability scores .#JAIC #RobustAI

⏳ Just 1 week to go! The ELLIS Doctoral Symposium #EDS2025 on #RobustAI kicks off in Warsaw on 25 August 2025. 🎓 Get ready for a week of cutting-edge talks, workshops, and networking with the AI research community! 👉 Check out the program here: elias-ai.eu/event/eds2025-…

Limited data input from sensors due to malfunctions or weather creates uncertainty .AI must work with sparse or noisy signals to make reliable predictions .#DataGaps #RobustAI

How deep models can stay stable on "messy" data #DataCleaning #RobustAI #TimeSeries #forecasting

🚀 Kicking off the week with our final #EDS2025 keynote speakers! 🎤Pau Rodríguez López (Apple) | @davsca1 (UZH & NASA JPL) | @ewa_szczurek (UW & Helmholtz Munich) Join us in Warsaw, Aug 25–29 🇵🇱 🗓️ Apply by July 31st! 🔗elias-ai.eu/event/eds2025-… #RobustAI #AIresearch #PhDlife

When your LiDAR sees ghosts👻 , who takes the blame? CoEvolution is shaping AI that not only works well, but fails well. 🤔What’s your strategy for adversarial sensor input in the wild? #AIsecurity #RobustAI #EdgeComputing #AutonomousSystems #EUFunded

Boosting LLM Robustness: Abstract Reasoning with AbstRaL for AI Researchers and Data Scientists #LanguageModels #RobustAI #AbstractReasoning #DataScience #ReinforcementLearning itinai.com/boosting-llm-r… Understanding the Importance of Robustness in Language Models Large languag…

2️⃣ Crowdsourced Adversarial Data @JoinSapien could reward contributors for generating challenging examples that trip up existing models, building datasets that harden AI against edge cases and exploits. #RobustAI #JoinSapien #deprem #Crypto

🎉 New paper accepted! @IITCNR 📄 “Robustness of decentralized learning to nodes and data disruption” 🔗 tinyurl.com/5a5e8fsj 💡 Decentralised learning can be surprisingly robust. Thanks to all the co-authors! #EdgeAI #FederatedLearning #RobustAI #ComputerCommunications

Reason 1: Models finally got good enough. Earlier agents spiraled into “context poisoning.” Now, LLMs recover from mistakes, handle retries, and keep workflows robust. MCP arrives right when model reliability hits the threshold. #RobustAI

Edge case: generate_commercial_asset("") → raises PromptError Input validation ensures quality + compliance. #RobustAI #EnterpriseCompliance

📢The #EDS2025 Keynote Lineup is here! Join us in Warsaw 🇵🇱 | Aug 25–29 for the EDS on #RobustAI - building AI that's reliable, safe & adaptable. 🎤With leading voices in AI & society, human-in-the-loop systems & applied ML: @SergeBelongie, @PetriMyllymaki, @nuriaoliver Join us!

Last but not least, presenting "ViUniT: Visual Unit Tests for More Robust Visual Programming" at #CVPR2025! 🗓️ Sun Jun 15, 10:30AM-12:30PM 📍 ExHall D Poster #346 🔗 Paper: arxiv.org/abs/2412.08859 📝 Blog: niebles.net/blog/2025/viun… #VisualProgramming #RobustAI

🔥dro is here: The First Comprehensive Python Package for Distributionally Robust ML! ✅ 79 DRO methods (14 formulations × 9 models) ✅ PyTorch/scikit-learn compatibility ✨ 10~1000x speedup via vectorization Check arxiv.org/pdf/2505.23565 #MachineLearning #RobustAI

创新剖析：@SaharalabsAl的'对抗训练'技术使AI系统在恶意攻击下仍保持稳定，这种鲁棒性对关键基础设施至关重要！#AIforALL #RobustAI

📢 Registration for ELLIS Doctoral Symposium #EDS2025 is open! Join us Aug 25-29 in Warsaw for a week on #RobustAI. Present your research, connect with experts, and network across Europe. Deadline: May 31, 2025 — register now! 🔗 elias-ai.eu/event/eds2025-… #ELIAS #HorizonEurope

Check out our NeurIPS 2022 paper MaskTune: Mitigating Spurious Correlations by Forcing to Explore 📄paper: i.sfu.ca/NASfOc 💻code: github.com/aliasgharkhani… @SFU_CompSci #deeplearning #computervision #robustAI #shortcuts

Use @rubrixml Metrics to: 1. Find inconsistent NER annotations & predictions 2. Do error analysis & compare different models: spaCy, FlairNLP, 🤗 3. Compare training with production datasets & across data slices rubrix.readthedocs.io/en/stable/guid… #nlproc #robustAI #python #opensource

⏳ Just 1 week to go! The ELLIS Doctoral Symposium #EDS2025 on #RobustAI kicks off in Warsaw on 25 August 2025. 🎓 Get ready for a week of cutting-edge talks, workshops, and networking with the AI research community! 👉 Check out the program here: elias-ai.eu/event/eds2025-…

"absolutely brilliant" —Nobel Laureate Danny Kahneman The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence Gary Marcus: arxiv.org/abs/2002.06177 #AIDebate #DeepLearning #RobustAI

🔥dro is here: The First Comprehensive Python Package for Distributionally Robust ML! ✅ 79 DRO methods (14 formulations × 9 models) ✅ PyTorch/scikit-learn compatibility ✨ 10~1000x speedup via vectorization Check arxiv.org/pdf/2505.23565 #MachineLearning #RobustAI

This event happens tomorrow: THURS 7/23! Register today to hear from FICO's @ScottZoldi to discuss #ResponsibleAI +#RobustAI + #AI as a disruptor and how using #alternatedata and timely #data with #AI can allow better decisions. bit.ly/30B2udE @EquifaxInsights

4/7 🛡️ The proposed method outperforms traditional metrics like BLEU and ROUGE by better reflecting the true quality of explanatory arguments. It’s also robust against adversarial attacks, ensuring reliability even in challenging cases. #RobustAI #AdversarialAI

TextFlint: Robustness Evaluation Toolkit for NLProc ➖ Multilingual toolkit for robust NLP evaluation ➖ 20 general & 60 task-specific transformations, adversarial attacks & subpopulations github.com/textflint/text… textflint.io/textflint #nlproc #robustAI #python #opensource

3/ Introducing 🛡️PROTECTED🛡️(Pre-training non-dominated Policies towards Online Adaptation): #RobustAI against more than worst-case attacks, optimizing robustness without compromising on everyday efficiency.

In 2021, the AI Asia Pacific Institute published the report 'Trustworthy Artificial Intelligence in the Asia-Pacific region'. Access it now via the link in our bio. #trustworthyai #robustai #ethicalai #aiapi

Up next at #NLPSummit Stage One: @malte_pietschfrom @deepset_ai presents 'Composable, Transparent, and Robust AI Applications.' Learn about building reliable AI solutions! #AI #RobustAI

LatticeFlow R&D team publishes a new state-of-the-art approach for improving the robustness of deep learning models without affecting their natural accuracy 🙌 NeurIPS paper 👉 lnkd.in/dx3V_7q3 Poster: 👉 lnkd.in/dTgyWQMA #NeurIPS2022 #robustness #robustai

The final element that is necessary to achieve trustworthy AI is a level of robustness. #trustworthyai #robustai #ethicalai #aiapi

AI Asia Pacific has released the trailer for the fourth season of it’s podcast. This season will focus on Asia-Pacific Collaboration and AI Governance in this region. Follow the link in our bio to listen. #trustworthyai #robustai #ethicalai #aiapi #aipodcast #spotify

We are proud to be a partner in the Europe-wide project @ai_humane , that is developing the foundations to shape the AI revolution in a direction that is beneficial to humans and adheres to European values and norms. fortiss.org/en/research/pr… #robustAI

"absolutely brilliant" —Nobel Laureate Danny Kahneman The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence Gary Marcus: lnkd.in/eFKe-Hf #AIDebate #DeepLearning #RobustAI

Research and development of state-of-the-art deepfake detection analytics with intuitive explanations and robustness to open-world variations as well as malicious adversarial examples. #adversarialexamples #deepfakedetection #robustai

👁️‍🗨️ Case study — @Privately_app reduced their real-world failures by 80%! Want to know how they did it? By using Lakera's MLTest platform. 👉 Read the full case study here: lakera.ai/insights/priva… #robustai #transparentai #computervision #mltesting #certification

📢The #EDS2025 Keynote Lineup is here! Join us in Warsaw 🇵🇱 | Aug 25–29 for the EDS on #RobustAI - building AI that's reliable, safe & adaptable. 🎤With leading voices in AI & society, human-in-the-loop systems & applied ML: @SergeBelongie, @PetriMyllymaki, @nuriaoliver Join us!

#RodneyBrooks Explains What #RobustAI Is Actually Doing dlvr.it/SVHq4n