Meta tags:
description=;
author=;

Headings (most frequently used words):

lecture, recordings, tutorials, documents, lectures, you, uva, deep, learning, ii, course, about, links, if, are, interested, in, older, versions, of, the, can, find, them, below, contact, us,

Text of the page (most frequently used words):
the (47), and (39), models (26), this (24), module (23), are (22), neural (20), you (18), networks (18), can (17), probabilistic (16), deep (15), lecture (13), learning (13), recordings (12), will (11), with (11), data (10), course (8), but (8), documents (8), dynamical (8), systems (8), physics (8), advanced (8), that (8), not (8), how (7), geometric (7), equivariant (7), also (6), lectures (6), tutorials (6), based (6), learn (6), distributions (6), these (6), nns (6), likelihood (6), estimation (6), group (6), uva (5), for (5), study (5), topic (5), such (5), their (5), part (5), general (5), cnns (5), symmetries (5), equivariance (5), see (4), model (4), autoregressive (4), bayesian (4), necessarily (4), techniques (4), amortised (4), background (4), generative (4), key (4), design (4), given (4), focus (4), probability (4), over (4), complex (4), structured (4), parameterise (4), parameters (4), gradient (4), modules (4), they (4), exclusive (4), vaes (4), upon (4), tool (4), cover (4), function (4), parameter (4), computational (4), convolutions (4), properties (4), gavves (3), knowledge (3), find (3), some (3), links (3), deadline (3), latent (3), way (3), into (3), many (3), most (3), very (3), spaces (3), via (3), all (3), chemistry (3), from (3), graph (3), methods (3), amsterdam (2), any (2), free (2), here (2), please (2), contact (2), uvadlc (2), interface (2), overlap (2), between (2), ordinary (2), partial (2), stochastic (2), differential (2), equations (2), where (2), found (2), implicit (2), functions (2), odes (2), used (2), like (2), navier (2), stokes (2), informed (2), well (2), fourier (2), inspired (2), architectures (2), don (2), need (2), content (2), understand (2), does (2), help (2), live (2), without (2), certain (2), variable (2), require (2), approximate (2), intractable (2), computations (2), principled (2), manner (2), those (2), discussed (2), variational (2), inference (2), because (2), concerns (2), thought (2), rather (2), special (2), instances (2), gives (2), ease (2), your (2), normalising (2), flows (2), energy (2), diffusion (2), processes (2), least (2), aspects (2), training (2), treatment (2), prescribe (2), sample (2), discrete (2), continuous (2), using (2), estimate (2), maximise (2), bounds (2), descent (2), goal (2), get (2), students (2), expand (2), toolbox (2), modelling (2), ideas (2), algorithms (2), compose (2), built (2), gaussians (2), one (2), type (2), another (2), reparameterisation (2), mle (2), etc (2), two (2), main (2), classes (2), depending (2), whether (2), assessed (2), tractably (2), set (2), observations (2), vector (2), view (2), byproduct (2), experiments (2), joint (2), observed (2), random (2), variables (2), however (2), may (2), perform (2), regularised (2), maximum (2), relationship (2), other (2), covers (2), touching (2), its (2), five (2), grids (2), groups (2), graphs (2), geodesics (2), gauges (2), strong (2), impact (2), made (2), fields (2), computer (2), vision (2), largely (2), attributed (2), fact (2), allow (2), weight (2), sharing (2), stability (2), dramatic (2), decrease (2), learnable (2), leveraging (2), architecture (2), enabling (2), arise (2), property (2), equip (2), efstratios (2), wilker (2), aziz (2), allows (2), problems (2), handle (2), core (2), convolutional (2), split (2), erik (2), bekkers (2), artificial (2), intelligence (2), liu (2), about (2), lab42, science, park, 904, 1098, netherlands, have, questions, recommendations, website, always, drop, line, should, feel, use, material, provided, kind, cite, case, instuctor, want, solutions, send, email, 2022, 2023, interested, older, versions, them, below, youtube, channel, canvas, datanose, piazza, useful, following, added, soon, third, representing, uncertainty, ferreira, second, https, uvagedl, github, recap, time, further, exploration, topics, covered, transformers, now, days, solved, sota, domains, derive, effectives, structure, presented, underlying, tools, context, state, art, steerable, introduce, class, rotational, flexible, powerful, successful, atomic, point, clouds, gauge, applicable, arbitrary, riemannian, manifolds, regular, basics, show, leverage, practical, accompanying, first, taught, msc, program, university, theory, namely, modern, multi, layered, trained, big, coordinated, teaching, assistants, tas, max, zhdanov, andrii, zadaianchuk, haochen, wang, philipp, tremuel, yongtuo, cong, metod, jazbec, alejandro, garcia, stefanos, achlatis, christos, athanasiadis, fereira, out, more, master, programme, old, materials, toggle, navigation,

Text of the page (random words):
uva deep learning ii course toggle navigation uva deep learning ii course about lectures tutorials old materials links contact uva deep learning ii course uva master s programme in artificial intelligence find out more about deep learning ii is taught in the msc program in artificial intelligence of the university of amsterdam in this course we study the theory of deep learning namely of modern multi layered neural networks trained on big data the course is coordinated by efstratios gavves erik bekkers wilker aziz fereira and christos athanasiadis the teaching assistants tas are stefanos achlatis alejandro garcia metod jazbec cong liu yongtuo liu philipp tremuel haochen wang andrii zadaianchuk max zhdanov lectures first part group equivariant deep learning erik bekkers this module covers the topic of geometric deep learning touching upon all its five g s grids groups graphs geodesics and gauges but with a strong focus on group equivariant deep learning the impact that cnns made in fields such as computer vision computational chemistry and physics can largely be attributed to the fact that convolutions allow for weight sharing geometric stability and a dramatic decrease in learnable parameters by leveraging symmetries in data and architecture design these enabling properties arise from the equivariance property of convolutions in this module you will learn how to equip neural networks with equivariance properties the module is split in 4 lectures with accompanying tutorials this module is split into 4 lectures lecture 1 regular group convolutional neural networks g cnns in this lecture we cover the basics of group convolutional nns and show how to leverage symmetries in data and practical problems lecture 2 steerable g cnns in this lecture we introduce a very general class of g cnns that allows to handle rotational symmetries in a flexible and powerful way these methods are at the core of the most successful methods to handle 3d data such as atomic point clouds but are also at the core of gauge equivariant methods that are applicable to arbitrary riemannian manifolds lecture 3 equivariant graph nns many problems in computational chemistry and computational physics are now a days solved via graph nns the sota in these domains derive their effectives from the geometric structure and symmetries presented by the data and underlying physics in this lecture cover tools for se 3 equivariance in the context of state of the art in geometric graph nns lecture 4 recap and or if time allows further exploration of topics covered in this module e g equivariant transformers geometric latent spaces â all the lectures can be find here https uvagedl github io documents no documents lecture recordings no recordings second part deep probabilistic models wilker aziz ferreira many if not most advanced dl models are probabilistic models or at the very least key aspects of their design and training are given probabilistic treatment the focus of this module or this part of the module is to learn to prescribe probability distributions over complex sample spaces discrete continuous structured parameterise these distributions using nns and estimate model parameters to maximise bounds on likelihood via gradient descent the goal is to get students to expand their toolbox to see modelling ideas and estimation algorithms as modules they can compose ie vi is not exclusive to vaes vaes are not necessarily built upon gaussians autoregressive models are not exclusive to one data type or another reparameterisation is a general tool mle is a general tool etc we cover two main classes of models depending on whether a key function the likelihood function can be assessed tractably given a set of observations and a parameter vector tl dr in this module you learn to view data as a byproduct of probabilistic experiments you will parameterise joint probability distributions over observed random variables however complex structured they may be and perform parameter estimation by regularised gradient based maximum likelihood estimation relationship to other modules advanced generative models are rather special instances of probabilistic models this module gives you some background knowledge that can ease your way into advanced generative models such as normalising flows energy based models and diffusion processes certain advanced probabilistic models e g latent variable models require techniques to approximate intractable computations in a principled manner those techniques are discussed in the amortised variational inference module because amortised vi concerns probabilistic models this module can be thought of as background to it bayesian models are also probabilistic but you don t necessarily need to content of this module to understand bayesian deep learning it does help but you can live without documents representing uncertainty in ml lecture recordings no recordings third part neural networks dynamical systems efstratios gavves in this module we will study the interface and overlap between neural networks dynamical systems ordinary partial stochastic differential equations and physics based neural networks we will study how and where dynamical systems be found in neural networks with implicit functions and neural odes we will also see how neural networks can be used to model dynamical systems like navier stokes with physics informed neural networks as well as with fourier inspired architectures and autoregressive neural networks documents no documents lecture recordings no recordings tutorials topic group equivariant deep learning deadline this module covers the topic of geometric deep learning touching upon all its five g s grids groups graphs geodesics and gauges but with a strong focus on group equivariant deep learning the impact that cnns made in fields such as computer vision computational chemistry and physics can largely be attributed to the fact that convolutions allow for weight sharing geometric stability and a dramatic decrease in learnable parameters by leveraging symmetries in data and architecture design these enabling properties arise from the equivariance property of convolutions in this module you will learn how to equip neural networks with equivariance properties tutorials no documents lecture recordings recordings will be added soon topic deep probabilistic models deadline many if not most advanced dl models are probabilistic models or at the very least key aspects of their design and training are given probabilistic treatment the focus of this module or this part of the module is to learn to prescribe probability distributions over complex sample spaces discrete continuous structured parameterise these distributions using nns and estimate model parameters to maximise bounds on likelihood via gradient descent the goal is to get students to expand their toolbox to see modelling ideas and estimation algorithms as modules they can compose ie vi is not exclusive to vaes vaes are not necessarily built upon gaussians autoregressive models are not exclusive to one data type or another reparameterisation is a general tool mle is a general tool etc we cover two main classes of models depending on whether a key function the likelihood function can be assessed tractably given a set of observations and a parameter vector tl dr in this module you learn to view data as a byproduct of probabilistic experiments you will parameterise joint probability distributions over observed random variables however complex structured they may be and perform parameter estimation by regularised gradient based maximum likelihood estimation relationship to other modules advanced generative models are rather special instances of probabilistic models this module gives you some background knowledge that can ease your way into advanced generative models such as normalising flows energy based models and diffusion processes certain advanced probabilistic models e g latent variable models require techniques to approximate intractable computations in a principled manner those techniques are discussed in the amortised variational inference module because amortised vi concerns probabilistic models this module can be thought of as background to it bayesian models are also probabilistic but you don t necessarily need to content of this module to understand bayesian deep learning it does help but you can live without tutorials no documents lecture recordings no recordings topic neural networks dynamical systems deadline in this module we will study the interface and overlap between neural networks dynamical systems ordinary partial stochastic differential equations and physics based neural networks we will study how and where dynamical systems be found in neural networks with implicit functions and neural odes we will also see how neural networks can be used to model dynamical systems like navier stokes with physics informed neural networks as well as with fourier inspired architectures and autoregressive neural networks tutorials no documents lecture recordings no recordings links some useful links for the course are the following piazza datanose canvas youtube channel if you are interested in older versions of the lectures you can find them below uvadlc 2023 uvadlc 2022 contact us if you have any questions or recommendations for the website or the course you can always drop us a line the knowledge should be free so feel also free to use any of the material provided here but please be so kind to cite us in case you are a course instuctor and you want the solutions please send us an email lab42 4 28 science park 904 1098 xh amsterdam the netherlands e gavves uva nl