波音游戏-波音娱乐城赌球打不开

Skip to main content

Self-interaction and learning in random structures

Professor Pierre Tarres
Date & Time
17 Apr 2019 (Wed) | 10:00 AM - 11:00 AM
Venue
B6605, Blue Zone, Floor 6, Yeung Kin Man Academic Building (AC1), CityU

Abstract

I will discuss self-interacting processes, and in particular the learning features that arise from that interaction. I will start with the reinforced random walk, introduced by Coppersmith and Diaconis in 1986, a nearest-neighbor walk on a discrete graph which reinforces on the number of times a vertex or an edge has been visited. The vertex-reinforced random walk displays strong localisation in the sense that it eventually gets stuck on a finite subset of vertices: I showed for instance in 2004 that on the integers it will eventually visit only five consecutive vertices infinitely often almost surely. On the other hand the edge-reinforced random walk surprisingly shows strikingly different behavior, and is explicitly linked to a supersymmetric hyperbolic sigma model in quantum field theory (Sabot and Tarrès, 2015), to a random Schr?dinger operator (Sabot, Tarrès and Zeng, 2017) and Dynkin's isomorphism (Sabot and Tarrès, 2016). It shows in particular a recurrence/transience phase transition in dimension 3 and higher (Disertori, Sabot and Tarrès, 2015), which settles a conjecture of Diaconis in 1986. I will also present some other related processes, in particular the strongly edge-reinforced random walk on which I proved with Limic in 2007 a conjecture of Sellke (1994) of eventual localisation on a single edge, and the so-called Brownian polymers, on which I proved two conjectures of Durrett and Rogers (1992), respectively with Mountford in 2008 and with Tóth and Valkó in 2012. Finally I will discuss applications in papers in collaboration on evolutionary biology questions, on the multi-armed bandit problem, on optimal bounds for online learning algorithms, and on reinforcement learning algorithms in game theory for language and network formation.

百家乐官网现金网排名| 百家乐赌博博彩赌博网| 百家乐官网的必赢术| 威尼斯人娱乐城金杯娱乐城| 百家乐视频二人麻将| 大连百家乐食品| 百家乐官网规则技法| 德州扑克技巧视频| 大发888娱乐城 34| 去澳门百家乐官网的玩法技巧和规则| 至尊百家乐官网20130301| 十大博彩网| 宝马会百家乐娱乐城| 打百家乐官网的技术| 菲律宾百家乐娱乐| 网上百家乐博彩正网| 金宝博网址| 百家乐娱乐平台开户| 百家乐现金网平台排行| 百家乐官网如何捕捉长龙| 百家乐官网老是输| 县级市| 百家乐官网网站出售| KK娱乐| 百家乐官网平注常赢玩法更 | 太子百家乐官网娱乐城| 百家乐官网改单软件| 泰盈娱乐城| 沙龙百家乐娱乐城| 永利百家乐游戏| 大佬百家乐官网现金网| 永利博百家乐的玩法技巧和规则| 太阳城娱乐城去大丰收娱乐| 芜湖市| 百家乐真钱在线| 龙腾国际娱乐| 加州百家乐官网的玩法技巧和规则 | 六合彩公式| A8百家乐官网娱乐场| 希尔顿百家乐官网娱乐城 | 百家乐官网轮盘桌|