Różnice

Różnice między wybraną wersją a wersją aktualną.

--- pl:start [2025/02/11 20:24] – [Luty] kkutt
+++ pl:start [2025/03/17 23:50] (aktualna) – [Luty] kkutt
@@ Linia 2: / Linia 2: @@
 ===== 2025 =====
+==== Marzec ====
+=== Zdalne seminarium PSSI i AI Bay: 13.03.2025 ===
+Polskie Stowarzyszenie Sztucznej Inteligencji (PSSI) serdecznie zaprasza na zorganizowane wspólnie z AI Bay seminarium na temat sztucznej inteligencji. Szczegóły podano poniżej.
+Kiedy: 13 marca 2025, 17.00-18.40 \\
+Gdzie: zdalnie, link będzie przesłany bezpośrednio przed wydarzeniem (wymagane wcześniejsze zgłoszenie) \\
+Zgłoszenia i dalsze informacje: https://aibay.ai/seminarium-ai-bay-13-03-2025/
+Seminarium będzie prowadzone w języku angielskim. \\
+Program wydarzenia podany jest poniżej.
+Serdecznie zapraszamy!
+Program:
+__17.00-17.50__ \\
+Who: **Karim Guirguis** \\
+What: **AI Agents: From Concept to Impact**
+**Bio:** Karim Guirguis is a senior AI solutions architect at SiMa.ai, specializing in AI applications for the industrial edge, robotics, and automotive industries. He got his Ph.D. in Deep Learning Computer Vision from KIT, Germany. At SiMa, he pairs deep technical domain expertise with an end-user point of view to help customers develop and deploy cutting-edge AI solutions. He brings extensive experience in helping organizations in various sectors gain a competitive advantage by utilizing AI to solve pressing pain points.
+__17.50-18.40__ \\
+Who: **Sebastian Cygert** \\
+What: **Robust fine-tuning with model merging.** \\
+**Bio:** Sebastian Cygert is a postdoc at Ideas NCBR and an assistant professor at the Gdańsk University of Technology, where he defended his PhD. Previously, he worked as an applied scientist at Amazon, contributing to projects such as the visual perception system for the autonomous Amazon Scout robot and speech synthesis for Alexa. In addition to this, he has extensive industry experience, including work on mathematical modeling for Moody’s Analytics and involvement in local startups. His research focuses on developing reliable and trustworthy machine learning models that can robustly adapt to dynamic environments.  He is a member of the ELLIS and his work has been published at leading AI conferences such as ECCV, NeurIPS, and ICLR. \\
+**Info:** Fine-tuning has become a standard practice in modern machine learning, enabling pre-trained models to adapt to specific tasks or domains. However, traditional fine-tuning often faces challenges such as instability, overfitting, and catastrophic forgetting, especially when dealing with limited data. Model merging offers a promising solution by integrating multiple fine-tuned models, ensuring that knowledge from diverse sources is preserved. In this work, we explore methods for merging fine-tuned models using our recent works.
+We'd like to invite you to join the discussion during the seminar.
 ==== Luty ====