ML Advanced Methods (אירוע)

אני שמח להזמין אתכם לאירוע הראשון שלנו בסדרת ML Advanced Methods שנעשים בשיתוף עם Mobileye. בכל אירוע מסוג זה, נסקור מספר עבודות בנושאים מתקדמים בתחום ה-ML ונציג פרקטיקות נפוצות לשימוש במודלים חזקים.  האירוע יתקיים ב-9.12 בשעה 18:00.

ההרצאות באירוע:

1. Asaf Bar Zvi (Lead Machine Learning Researcher at Mobileye) – ME NSS – Semantic Segmentation Network.

NSS takes the ability of deep neural networks to the limit with much higher performance level than achieving on academic benchmarks, using architecture novelties and smart data usage.

2. Erez Farhan (Research Scientist at BeyondMinds) – Self-supervised learning in computer-vision (recent progress)

Representation learning is arguably what deep-learning is all about. It has been shown that given enough manual annotation for a given task, effective visual representations can be learned and even be successfully transferred to related tasks with significantly fewer labels. Rapidly expanding to new tasks and data domains, manual labeling can no longer stand the pace. Accordingly, fascinating efforts and breakthroughs are being made for allowing effective

representation learning without human supervision. This talk will cover recent efforts in self-supervised representation learning, where the labels come from the data itself!. We will discuss the implications, underlying assumptions, and possible information-theoretic pitfalls of these recent ideas.