Stage-based Hyper-parameter Optimization for Deep Learning. (arXiv:1911.10504v1 [cs.LG])

As deep learning techniques advance more than ever, hyper-parameter optimization is the new major workload in deep learning clusters. Although hyper-parameter optimization is crucial in training deep learning models for high model performance, effectively executing such a computation-heavy workload still remains a challenge. We observe that numerous trials issued from existing hyper-parameter optimization algorithms share…

dpVAEs: Fixing Sample Generation for Regularized VAEs. (arXiv:1911.10506v1 [cs.LG])

Unsupervised representation learning via generative modeling is a staple to many computer vision applications in the absence of labeled data. Variational Autoencoders (VAEs) are powerful generative models that learn representations useful for data generation. However, due to inherent challenges in the training objective, VAEs fail to learn useful representations amenable for downstream tasks. Regularization-based methods…

Semi-Supervised Hierarchical Recurrent Graph Neural Network for City-Wide Parking Availability Prediction. (arXiv:1911.10516v1 [cs.LG])

The ability to predict city-wide parking availability is crucial for the successful development of Parking Guidance and Information (PGI) systems. Indeed, the effective prediction of city-wide parking availability can improve parking efficiency, help urban planning, and ultimately alleviate city congestion. However, it is a non-trivial task for predicting citywide parking availability because of three major…

The route to chaos in routing games: When is Price of Anarchy too optimistic?. (arXiv:1906.02486v2 [cs.GT] UPDATED)

Routing games are amongst the most studied classes of games. Their two most well-known properties are that learning dynamics converge to equilibria and that all equilibria are approximately optimal. In this work, we perform a stress test for these classic results by studying the ubiquitous dynamics, Multiplicative Weights Update, in different classes of congestion games,…

Why so many significant phase III results in clinical trials?. (arXiv:1907.00185v2 [econ.GN] UPDATED)

Planning and execution of clinical research and publication of results should conform to the highest ethical standards, given that human lives are at stake. However, economic incentives can generate conflicts of interest for investigators, who may be inclined to withhold unfavorable results or even tamper with the data. Analyzing p-values reported to the ClinicalTrials.gov registry…

Risk-dependent centrality in economic and financial networks. (arXiv:1907.07908v3 [q-fin.MF] UPDATED)

Node centrality is one of the most important and widely used concepts in the study of complex networks. Here, we extend the paradigm of node centrality in financial and economic networks to consider the changes of node “importance” produced not only by the variation of the topology of the system but also as a consequence…

Collectivised Pension Investment. (arXiv:1909.12730v2 [q-fin.PM] UPDATED)

We study the optimal management of a collectivised pension fund, where all investors agree that the assets of deceased members are shared among the survivors. We find that for realistic parameters based on the UK pensions market, a collectivised fund achieves an approximately 20% better return than either an annuity or a personal investment fund.…

Scheduling and control over networks using MPC with time-varying terminal ingredients. (arXiv:1911.10780v1 [eess.SY])

Rollout approaches are an effective tool to address the problem of co-designing the transmission schedule and the corresponding input values, when the controller is connected to the plant via a resource-constrained communication network. These approaches typically employ an MPC, activated at multiples of the period length of a base transmission schedule. Using multi-step invariant terminal…

Vol-of-vol expansion for (rough) stochastic volatility models. (arXiv:1910.03245v3 [math.PR] UPDATED)

We introduce an asymptotic small noise expansion, a so called vol-of-vol expansion, for potentially infinite dimensional and rough stochastic volatility models. Thereby we extend the scope of existing results for finite dimensional models and validate claims for infinite dimensional models. Furthermore we provide new, explicit (in the sense of non-recursive) representations of the so-called push-down…