Technion – Israel Institute of Technology

The tremendous success of the Machine Learning paradigm heavily relies on the development of powerful optimization methods, and the canonical algorithm for training learning models is SGD (Stochastic Gradient Descent). Nevertheless, the latter is quite different from Gradient Descent (GD) which is its noiseless counterpart. Concretely, SGD requires a careful choice of the learning rate, which relies on the properties of the noise as well as the quality of initialization.

 It further requires the use of a test set to estimate the generalization error throughout its run. In this talk, we will present a new SGD variant that obtains the same optimal rates as SGD, while using noiseless machinery as in GD. Concretely, it enables to use the same fixed learning rate as GD and does not require to employ a test/validation set. Curiously, our results rely on a novel gradient estimate that combines two recent mechanisms which are related to the notion of momentum.

Finally, as much as time permits, I will discuss several applications where our method can be extended.

The usual language of algebraic geometry is not appropriate for Arithmetical geometry: addition is singular at the real prime. We developed two languages that overcome this problem: one replace rings by the collection of “vectors” or by bi-operads and another based on “matrices” or props. These are the two languages of [Har17], but we omit the involutions which brings considerable simplifications. Once one understands the delicate commutativity condition one can proceed following Grothendieck footsteps exactly. The square matrices, when viewed up to conjugation, give us new commutative rings with Frobenius endomorphisms.

Connectivity and percolation are two well studied phenomena in random graphs. 

In this talk we will discuss higher-dimensional analogues of connectivity and percolation that occur in random simplicial complexes.

Simplicial complexes are a natural generalization of graphs that consist of vertices, edges, triangles, tetrahedra, and higher dimensional simplexes.

We will mainly focus on random geometric complexes. These complexes are generated by taking the vertices to be a random point process, and adding simplexes according to their geometric configuration.

Our generalized notions of connectivity and percolation use the language of homology - an algebraic-topological structure representing cycles of different dimensions.

In this talk we will discuss recent results analyzing phase transitions related to these topological phenomena. 

It is a classical theorem of Magnus that the word problem for one-relator groups is solvable; its precise complexity remains unknown. A geometric characterization of the complexity is given by the Dehn function. I will present joint work with Daniel Woodhouse showing that one-relator groups have a rich collection of Dehn functions, including the Brady--Bridson snowflake groups on which our work relies.
 

It is a classical theorem of Magnus that the word problem for one-relator groups is solvable; its precise complexity remains unknown. A geometric characterization of the complexity is given by the Dehn function. I will present joint work with Daniel Woodhouse showing that one-relator groups have a rich collection of Dehn functions, including the Brady--Bridson snowflake groups on which our work relies.