#foundationmodels — Public Fediverse posts

Generalist AI veröffentlicht das multimodale Robotik-Modell GEN-1 mit 99 Prozent Erfolgsquote bei physischen Aufgaben.

Die Architektur verzichtet auf Teleoperation im Pre-Training und nutzt 500.000 Stunden Sensor-Daten menschlicher Hände für spontane Echtzeit-Fehlerkorrektur. Eine Adaption an neue Hardware benötigt nur eine Stunde spezifischer Roboterdaten.

#GeneralistAI #Robotik #FoundationModels #LLM #News
https://www.all-ai.de/news/news26/gen-1-generalist-new

The Model That Knows Too Much: How AI Can Leak What It Learned
https://youtu.be/pwA5nASJpoo #Cybersecurity #ArtificialIntelligence #AIsecurity #ModelInversionAttacks #DataPrivacy #MachineLearning #AISafety #Infosec #DigitalRisk #FoundationModels #DataProtection #CyberRisk

Five Architectures for Time Series Forecasting with Large Language Models
Large Language Models are increasingly being applied to time series forecasting. Not as chatbots, but as prediction engines that leverage the pattern recognition capabilitie
https://www.hylkerozema.nl/2026/02/25/five-architectures-for-time-series-forecasting-with-large-language-models/
#DataScience #MachineLearningEngineering #DataScience #Forecasting #FoundationModels #LLM #MachineLearning #TimeSeries #Transformer

Five Architectures for Time Series Forecasting with Large Language Models
Large Language Models are increasingly being applied to time series forecasting. Not as chatbots, but as prediction engines that leverage the pattern recognition capabilitie
https://www.hylkerozema.nl/2026/02/25/five-architectures-for-time-series-forecasting-with-large-language-models/
#DataScience #MachineLearningEngineering #DataScience #Forecasting #FoundationModels #LLM #MachineLearning #TimeSeries #Transformer

Five Architectures for Time Series Forecasting with Large Language Models
Large Language Models are increasingly being applied to time series forecasting. Not as chatbots, but as prediction engines that leverage the pattern recognition capabilitie
https://www.hylkerozema.nl/2026/02/25/five-architectures-for-time-series-forecasting-with-large-language-models/
#DataScience #MachineLearningEngineering #DataScience #Forecasting #FoundationModels #LLM #MachineLearning #TimeSeries #Transformer

Five Architectures for Time Series Forecasting with Large Language Models
Large Language Models are increasingly being applied to time series forecasting. Not as chatbots, but as prediction engines that leverage the pattern recognition capabilitie
https://www.hylkerozema.nl/2026/02/25/five-architectures-for-time-series-forecasting-with-large-language-models/
#DataScience #MachineLearningEngineering #DataScience #Forecasting #FoundationModels #LLM #MachineLearning #TimeSeries #Transformer

Five Architectures for Time Series Forecasting with Large Language Models
Large Language Models are increasingly being applied to time series forecasting. Not as chatbots, but as prediction engines that leverage the pattern recognition capabilitie
https://www.hylkerozema.nl/2026/02/25/five-architectures-for-time-series-forecasting-with-large-language-models/
#DataScience #MachineLearningEngineering #DataScience #Forecasting #FoundationModels #LLM #MachineLearning #TimeSeries #Transformer

Europe’s competitiveness will depend on linking its foundation models with industrial applicability, governance standards, and sectoral integration. SOOFI: ~100B parameter, reasoning model and agentic applications. #SovereignAI #EuropeanAI #FoundationModels

SOOFI Launches: Europe’s Path ...

Europas Wettbewerbsfähigkeit hängt künftig davon ab, eigene Foundation-Modelle mit industrieller Anwendbarkeit, Governance-Standards und sektoraler Integration zu verbinden.
Mit SOOFI startet eine der ambitioniertesten europäischen Initiativen für souveräne Open-Source-Foundation-Modelle – geplant ist ein Sprachmodell mit rund 100 Mrd. Parametern, ergänzt durch ein spezialisiertes Reasoning-Modell und agentische Anwendungen.
https://www.l3s.de/soofi-launches-europes-path-towards-its-own-ai-language-models/
#SovereignAI #EuropeanAI #FoundationModels

Data Poisoning — The Silent Sabotage of AI
https://youtu.be/J-tsemViDXk #Cybersecurity #ArtificialIntelligence #AIsecurity #DataPoisoning #MachineLearning #AIrisk #AISafety #ModelSecurity #FoundationModels #CyberRisk #Infosec #DigitalTrust

Nice set of embodied foundation models this week:

in 2B, 8B, and 30B variants + RynnBrain‑Plan (manipulation planning), RynnBrain‑Nav (navigation), and RynnBrain‑CoP (spatial reasoning).

https://alibaba-damo-academy.github.io/RynnBrain.github.io/

ABot-M0: VLA Foundation Model for Robotic Manipulation, https://github.com/amap-cvlab/ABot-Manipulation

ABot-N0: Unified VLA for embodied navigation, https://amap-cvlab.github.io/ABot-Navigation/ABot-N0/

#foundationModels #embodied #robotics #robotFoundationModels #embodiedFoundationModels #VLA

via #AIFoundry : Beyond the Prompt – Why and How to Fine-tune Your Own Models

https://ift.tt/DpcF2h4
#finetuning #microsoftfoundry #azureai #enterpriseai #llm #languagemodels #foundationmodels #promptengineering #rag #retrievalaugmentedgeneration #sft #supervisedfinetuning #r…

Meeting Minutes: Pragmatic AI: Foundation models running entirely on-device. Apple's exclusive iOS 26 AI. No other platform can do this.
#FoundationModels #iOS26 #Exclusive hashtag#Simple #DataPrivacy
#OnDevice #iOS26 #Privacy #Productivity
#Focus #Minimalism #AI #Private.

https://apps.apple.com/gb/app/meeting-minutes-pragmatic-ai/id6752467167

Today, the #FoundationModels (Transformer, LLM, etc.) of #AI predominate every field of human endeavour. Perhaps a grand-sounding moniker, like the following, is appropriate for this much-revered technology:

Regnal Transformative Foundation Models

And recent studies in various fields had shown conclusively that quick access to #information in the short-term erodes the drive to acquire and retain #knowledge for the long-term. So, this estimable label's #RTFM acronym, which harkens back to that famous, 1980s tech aphorism, is equally appropriate.

Die @Cyberagentur startet HEGEMON, einen europaweit einzigartigen Forschungswettbewerb zur Bewertung und Anpassung von Foundation Models für sicherheitskritische Anwendungen. Vier Teams entwickeln Benchmarks und KI-Modelle für komplexe Aufgaben im Geoinformationswesen.
Mehr dazu: https://t1p.de/7ct97
#Cyberagentur #HEGEMON #KI #FoundationModels #Cybersicherheit #Benchmarking

HEGEMON startet: Vier Teams entwickeln Benchmarks und KI-Modelle für sicherheitskritische Anwendungen. Mehr zum Wettbewerb: https://t1p.de/7ct97
#HEGEMON #Cyberagentur #KI #FoundationModels #Cybersicherheit
https://nachrichten.idw-online.de/2025/12/10/europas-weg-zu-einer-transparenteren-anwendung-von-foundation-models

Computational Biology Lead
Bioptimus

Bioptimus is building the first universal AI foundation model for biology to fuel breakthrough discoveries and accelerate innovation in biomedicine.

See the full job description on jobRxiv: https://jobrxiv.org/job/bioptimus-27778-computational-biology-lead/

#bioinformatics #computationalbiology #foundationmodels #multiomics #ScienceJobs #hiring #research
https://jobrxiv.org/job/bioptimus-27778-computational-biology-lead/?fsp_sid=5086

#GenerativeAI, #FoundationModels, #LLMs, and all of that hokey nonsense shall not appear in my #robotics roadmaps as anything other than a neat research item until it can demonstrate a feasible path to #FunctionalSafety or mathematical completeness.

I lead #Product on the largest mobile-#robotic fleet known to humankind. I will not entrust decisions that could maim or kill to a pile of nondeterminate math prone to “hallucinations” or confabulation.

#ProductManagement

#GenerativeAI, #FoundationModels, #LLMs, and all of that hokey nonsense shall not appear in my #robotics roadmaps as anything other than a neat research item until it can demonstrate a feasible path to #FunctionalSafety or mathematical completeness.

I lead #Product on the largest mobile-#robotic fleet known to humankind. I will not entrust decisions that could maim or kill to a pile of nondeterminate math prone to “hallucinations” or confabulation.

#ProductManagement

#GenerativeAI, #FoundationModels, #LLMs, and all of that hokey nonsense shall not appear in my #robotics roadmaps as anything other than a neat research item until it can demonstrate a feasible path to #FunctionalSafety or mathematical completeness.

I lead #Product on the largest mobile-#robotic fleet known to humankind. I will not entrust decisions that could maim or kill to a pile of nondeterminate math prone to “hallucinations” or confabulation.

#ProductManagement

#GenerativeAI, #FoundationModels, #LLMs, and all of that hokey nonsense shall not appear in my #robotics roadmaps as anything other than a neat research item until it can demonstrate a feasible path to #FunctionalSafety or mathematical completeness.

I lead #Product on the largest mobile-#robotic fleet known to humankind. I will not entrust decisions that could maim or kill to a pile of nondeterminate math prone to “hallucinations” or confabulation.

#ProductManagement

AnyLoc: Towards Universal Visual Place Recognition

#SelfPromotion

Using features from intermediate blocks of foundation models like DINO and DINOv2, combined with conventional aggregation methods, can yield good descriptors for visual place recognition in diverse environments. No training or fine-tuning.

Summary: https://huggingface.co/papers/2308.00688#64cdbc0fe13e7c2bc5775aaa
arXiv: https://arxiv.org/abs/2308.00688
Links: [Website](https://anyloc.github.io/), [PapersWithCode](https://paperswithcode.com/paper/anyloc-towards-universal-visual-place)

#FoundationModels #VPR #arXiv