Format results
Gravitational Laboratories for Nuclear Physics (in light of GWTC-2)
Reed Essick University of Toronto
Special Topics in Astrophysics - Numerical Hydrodynamics - Lecture 14
Daniel Siegel University of Greifswald
New physics in flat Moire bands
Erez Berg Weizmann Institute of Science
What Are the Statistical Limits of Offline Reinforcement Learning With Function Approximation?
Sham Kakade (University of Washington & Microsoft Research)An Alternative Softmax Operator for Reinforcement Learning
Michael Littman (Brown University)Contextuality-by-default for behaviours in compatibility scenarios
Alisson Cordeiro Alves Tezzin Universidade Estadual Paulista (UNESP)
On the Global Convergence and Approximation Benefits of Policy Gradient Methods
Daniel Russo (Columbia University)Corruption Robust Exploration in Episodic Reinforcement Learning
Aleksandrs Slivkins (Microsoft Research NYC)Representation Learning and Exploration in Reinforcement Learning
Akshay Krishnamurthy (Microsoft Research)
Strategies for solving the Fermi-Hubbard model on near-term quantum computers
Ashley Montanaro University of Bristol
The Fermi-Hubbard model is of fundamental importance in condensed-matter physics, yet is extremely challenging to solve numerically. Finding the ground state of the Hubbard model using variational methods has been predicted to be one of the first applications of near-term quantum computers. In this talk, I will discuss recent work which carried out a detailed analysis and optimisation of the complexity of variational quantum algorithms for finding the ground state of the Hubbard model, including extensive numerical experiments for systems with up to 12 sites. The depth complexities we find are substantially lower than previous work. If our numerical results on small lattice sizes are representative of the somewhat larger lattices accessible to near-term quantum hardware, they suggest that optimising over quantum circuits with a gate depth less than a thousand could be sufficient to solve instances of the Hubbard model beyond the capacity of classical exact diagonalisation. I will also discuss a proof-of-principle implementation on Rigetti quantum computing hardware.
The talk is based on joint work with Chris Cade, Lana Mineh and Stasja Stanisic ( arXiv:1912.06007 , arXiv:2006.01179 ).Gravitational Laboratories for Nuclear Physics (in light of GWTC-2)
Reed Essick University of Toronto
Gravitational waves provide a unique way to study the universe. From the initial direct detection of coalescing black holes in 2015, to the ground-breaking multimessenger observations of coalescing neutron stars in 2017, and continuing with the now routine detection of merging stellar remnants, gravitational wave astronomy has quickly matured into a key aspect of modern physics. After briefly discussing what we've begun to learn from the new gravitational-wave transient catalog published by the LIGO, Virgo, and KAGRA collaborations (GWTC-2), I will discuss novel tests of fundamental physics GWs enable. In particular, I will focus on our current understanding of matter effects during the inspiral of compact binaries and matter at supranuclear densities, including possible phase transitions, through tests of neutron star structure. Detailed knowledge of dynamical interactions between coalescing stars, observations of extreme relativistic astrophysical systems, terrestrial experiments, and nuclear theory provide complementary views of fundamental physics. I will show how combining aspects from all these will improve our understanding of dense matter through the example of how we can determine whether newly observed objects are neutron stars or black holes.
Special Topics in Astrophysics - Numerical Hydrodynamics - Lecture 14
Daniel Siegel University of Greifswald
The Connected Universe: Relating Early, Intermediate and Late Universe with cosmological data
Vivian Miranda University of Arizona
The standard model of cosmology is built upon on a series of propositions on how the early, intermediate, and late epochs of the Universe behave. In particular, it predicts that dark energy and dark matter currently pervades the cosmos. Understanding the properties of the dark sector is plausibly the biggest challenge in theoretical physics. There is, however, a broad assumption in cosmology that the Universe on its earlier stages is fully understood and that discrepancies between the standard model of cosmology and current data are suggestive of distinct dark energy properties. Uncertainties on the validity of this hypothesis are not usually taken into account when forecasting survey capabilities, even though our investigations might be obfuscated if the intermediate and early Universe did behave abnormally. In this colloquium, I propose a program to investigate dark energy and earlier aspects of our Universe simultaneously, through space missions in the 2020s in combination with ground-based observatories. This program will help guide the strategy for the future LSST and WFIRST supernovae and weak lensing surveys. My investigations on how properties of the early and intermediate Universe affect inferences on dark energy (and vice-versa) will also support community understanding of how future missions can be employed to test some of the core hypotheses of the standard model of cosmology.
New physics in flat Moire bands
Erez Berg Weizmann Institute of Science
Flat bands in Moire superlattices are emerging as a fascinating new playground for correlated electron physics. I will present the results of several studies inspired by these developments. First, I will address the question of whether superconductivity is possible even in the limit of a perfectly flat band. Then, I will discuss transport properties of a spin-polarized superconductor in the limit of zero spin-orbit coupling, where the topological structure of the order parameter space allows for a new dissipation mechanism not known from conventional superconductors. If time allows, I will also discuss the interpretation of new measurements of the electronic compressibility in twisted bilayer graphene, indicating a cascade of symmetry-breaking transitions as a function of the density of carriers in the system.
References:
https://arxiv.org/abs/2006.10073
Measuring neutrino oscillations with IceCube and beyond
Juan Pablo YanezNeutrino oscillations have been probed during the last few decades using multiple neutrino sources and experimental set-ups. In recent years, very large volume neutrino telescopes have started contributing to the field. These large and sparsely instrumented detectors observe atmospheric neutrinos for combinations of baselines and energies inaccessible to other experiments. IceCube, the largest neutrino telescope in operation, has used this to measure standard oscillations and place limits on exotic proposals, such as sterile neutrinos. In this talk, I will go over the newest results from IceCube as well as the improvements expected thanks to a new detector upgrade to be deployed in the near future.
What Are the Statistical Limits of Offline Reinforcement Learning With Function Approximation?
Sham Kakade (University of Washington & Microsoft Research)The area of offline reinforcement learning seeks to utilize offline (observational) data to guide the learning of (causal) sequential decision making strategies. The hope is that offline reinforcement learning coupled with function approximation methods (to deal with the curse of dimensionality) can provide a means to help alleviate the excessive sample complexity burden in modern sequential decision making problems. As such, the approach is becoming increasingly important to numerous areas of science, engineering, and technology. However, the extent to which this broader approach can be effective is not well understood, where the literature largely consists of sufficient conditions. This work focuses on the basic question of what are necessary representational and distributional conditions that permit provable sample-efficient offline reinforcement learning. Perhaps surprisingly, our main result shows that even if: i) we have realizability in that the true value function of _every_ policy is linear in a given set of features and 2) our off-policy data has good coverage over all features (under a strong spectral condition), then any algorithm still (information-theoretically) requires a number of offline samples that is exponential in the problem horizon in order to non-trivially estimate the value of _any_ given policy. Our results highlight that sample-efficient, offline policy evaluation is simply not possible unless significantly stronger conditions hold; such conditions include either having low distribution shift (where the offline data distribution is close to the distribution of the policy to be evaluated) or significantly stronger representational conditions (beyond realizability). This is joint work with Ruosong Wang and Dean Foster.An Alternative Softmax Operator for Reinforcement Learning
Michael Littman (Brown University)A softmax operator applied to a set of values acts somewhat like the maximization function and somewhat like an average. In sequential decision making, softmax is often used in settings where it is necessary to maximize utility but also to hedge against problems that arise from putting all of one's weight behind a single maximum utility decision. The Boltzmann softmax operator is the most commonly used softmax operator in this setting, but we show that this operator is prone to misbehavior. In this work, we study a differentiable softmax operator that, among other properties, is a non-expansion ensuring a convergent behavior in learning and planning. We introduce a variant of SARSA algorithm that, by utilizing the new operator, computes a Boltzmann policy with a state-dependent temperature parameter. We show that the algorithm is convergent and that it performs favorably in practice. (With Kavosh Asadi.)Contextuality-by-default for behaviours in compatibility scenarios
Alisson Cordeiro Alves Tezzin Universidade Estadual Paulista (UNESP)
The compatibility-hypergraph approach to contextuality (CA) and the contextuality-by-default approach (CbD) are usually presented as products of entirely different views on how physical measurements and measurement contexts should be understood: the latter is based on the idea that a physical measurement has to be seen by a collection of random variables, one for each context containing that measurement, while the imposition of the non-disturbance condition as a physical requirement in the former precludes such interpretation of measurements. The aim of our work is to present both approaches as entirely compatible ones and to introduce in the compatibility-hypergraph approach ideas which arises from contextuality-by-default. We
introduce in CA the non-degeneracy condition, which is the analogous of consistent connectedness (an important concept from CbD), and prove that this condition is, in general, weaker than non-disturbance. The set of non-degenerate behaviours defines a polytope, therefore one can characterize non-degeneracy using a finite set of linear inequalities. We introduce extended contextuality for behaviours and prove that a behaviour is non-contextual in the standard sense if and only if it is non-degenerate and non-contextual in the extended sense. Finally, we use extended scenarios and behaviours to shed new light on our results.
On the Global Convergence and Approximation Benefits of Policy Gradient Methods
Daniel Russo (Columbia University)Policy gradients methods apply to complex, poorly understood, control problems by performing stochastic gradient descent over a parameterized class of polices. Unfortunately, due to the multi-period nature of the objective, policy gradient algorithms face non-convex optimization problems and can get stuck in suboptimal local minima even for extremely simple problems. This talk with discus structural properties – shared by several canonical control problems – that guarantee the policy gradient objective function has no suboptimal stationary points despite being non-convex. Time permitting, I’ll then zoom in on the special case of state aggregated policies and a proof showing that policy gradient converges to better policies than its relative, approximate policy iteration.Corruption Robust Exploration in Episodic Reinforcement Learning
Aleksandrs Slivkins (Microsoft Research NYC)We initiate the study of episodic RL under adversarial corruptions in both the rewards and the transition probabilities of the underlying system. Our solution adapts to unknown level of corruption, degrading gracefully in the total corruption encountered. In particular, we attain near-optimal regret for a constant level of corruption. We derive results for "tabular" MDPs as well as MDPs that admit a linear representation. Notably, we provide the first sublinear regret guarantee that goes beyond i.i.d. transitions in the bandit-feedback model for episodic RL. We build on a new framework which combines the paradigms of "optimism under uncertainty" and "successive elimination". Neither paradigm alone suffices: "optimism under uncertainty", common in the current work on stochastic RL, cannot handle corruptions, while "successive elimination" works for bandits with corruptions, but is provably inefficient even for stochastic RL. Joint work Thodoris Lykouris, Max Simchowitz and Wen Sun https://arxiv.org/abs/1911.08689Representation Learning and Exploration in Reinforcement Learning
Akshay Krishnamurthy (Microsoft Research)I will discuss new provably efficient algorithms for reinforcement in rich observation environments with arbitrarily large state spaces. Both algorithms operate by learning succinct representations of the environment, which they use in an exploration module to acquire new information. The first algorithm, called Homer, operates in a block MDP model and uses a contrastive learning objective to learn the representation. On the other hand, the second algorithm, called FLAMBE, operates in a much richer class of low rank MDPs and is model based. Both algorithms accommodate nonlinear function approximation and enjoy provable sample and computational efficiency guarantees.