Astrophysics and Relativity

Name	Category	Phone	Office
Asida Shimon	Faculty	02 6586 854	Popick 115A
Bach Yossi	Graduate Students
Balberg Shmuel	Faculty	02 6586 851	Popick 112
Barkat Zalman	Emeriti	02 6586 857	Popick 120
Benyamin David	Post-Docs		Kaplun 100
Birnboim Yuval	Faculty	02 6586 862	Popick Bldg.
Braun Arie	Senior Scientists	02 6584 417	Popick 122
Brumberg Itai	Graduate Students
Chan Chi Ho (Edwin)	Post-Docs	02 6584 997	Kaplun 101
Cornuault Nicolas	Post-Docs

Pages

Upcoming Events

Yesterday

Nonlinear Physics seminar: "Statistical mechanics of wide but finite artificial neural networks"
- Start time: 12:00pm
- End date: Wednesday, December 6th
- End time: 01:30pm
- Lecturer: Mr. Gadi Naveh (HUJI)Abstract: Over the past decade, Deep neural networks (DNNs) have brought significant advancements to various fields of science and technology. Key to their success is feature learning, enabling them to extract useful representations from raw data. However, DNNs remain challenging to interpret and lack a comprehensive analytical understanding. Recent studies revealed that DNNs in the infinite width limit are equivalent to Gaussian Processes (GPs), which are well-understoodanalytically but lack feature learning capabilities due to pre-determined features.In this thesis, we aim to bridge the gap between GPs and finite DNNs, using approaches inspired by theoretical physics. Employing statistical field theory, we treat the DNN output as a random object from a statistical ensemble. We performed extensive computer simulations of DNNs trained in both stylized and real-world settings. As for analytical methods, we first employ a perturbative approach to obtain the leading correction in 1/width to the GP limit. Subsequently, we develop a self-consistent mean-field type theory, providing a GP description of finite DNNs with a data-dependent kernel that embodies feature learning properties.
- View this event in Google Calendar
Position at Kyoto
- Start time: 02:00pm
- End date: Wednesday, December 6th
- End time: 03:00pm
- View this event in Google Calendar

Search form

Astrophysics and Relativity

Pages