Reinforcement learning (RL) is a technique to learn the control policy for an agent that interacts with a stochastic environment. In any given state, the agent takes some action, and the environment determines the probability distribution over the ne
xt state as well as gives the agent some reward. Most RL algorithms typically assume that the environment satisfies Markov assumptions (i.e. the probability distribution over the next state depends only on the current state). In this paper, we propose a model-based RL technique for a system that has non-Markovian dynamics. Such environments are common in many real-world applications such as in human physiology, biological systems, material science, and population dynamics. Model-based RL (MBRL) techniques typically try to simultaneously learn a model of the environment from the data, as well as try to identify an optimal policy for the learned model. We propose a technique where the non-Markovianity of the system is modeled through a fractional dynamical system. We show that we can quantify the difference in the performance of an MBRL algorithm that uses bounded horizon model predictive control from the optimal policy. Finally, we demonstrate our proposed framework on a pharmacokinetic model of human blood glucose dynamics and show that our fractional models can capture distant correlations on real-world datasets.
Intrinsically broken symmetries in the bulk of topological insulators (TIs) are manifested in their surface states. In spite of particle-hole asymmetry in TIs, it has often been assumed that their surface states are characterized by a particle-hole s
ymmetric Dirac energy dispersion. In this work we demonstrate that the effect of particle-hole asymmetry is essential to correctly describe the energy spectrum and the magneto-optical response in TIs thin-films. In thin-films of TIs with a substantial degree of particle-hole symmetry breaking, such as Sb$_2$Te$_3$, the longitudinal optical conductivity displays absorption peaks arising from optical transitions between bulk and surface Landau levels for low photon energies. The transition energies between the bulk and surface Landau levels exhibit clearly discernable signatures from those between surface Landau levels due to their distinct magnetic field dependence. Bulk contributions to the magneto-optical conductivity in a TI thin-film are enhanced via one type of doping while being suppressed by the other. This asymmetric dependence on type of doping aids in revealing the particle-hole asymmetry in TI thin-films.
Neural models have transformed the fundamental information retrieval problem of mapping a query to a giant set of items. However, the need for efficient and low latency inference forces the community to reconsider efficient approximate near-neighbor
search in the item space. To this end, learning to index is gaining much interest in recent times. Methods have to trade between obtaining high accuracy while maintaining load balance and scalability in distributed settings. We propose a novel approach called IRLI (pronounced `early), which iteratively partitions the items by learning the relevant buckets directly from the query-item relevance data. Furthermore, IRLI employs a superior power-of-$k$-choices based load balancing strategy. We mathematically show that IRLI retrieves the correct item with high probability under very natural assumptions and provides superior load balancing. IRLI surpasses the best baselines precision on multi-label classification while being $5x$ faster on inference. For near-neighbor search tasks, the same method outperforms the state-of-the-art Learned Hashing approach NeuralLSH by requiring only ~ {1/6}^th of the candidates for the same recall. IRLI is both data and model parallel, making it ideal for distributed GPU implementation. We demonstrate this advantage by indexing 100 million dense vectors and surpassing the popular FAISS library by >10% on recall.
We report the first measurement of proper motions in the SN1006 remnant (G327.6+14.6) based entirely on digital images. CCD images from three epochs spanning a period of 11 years are used: 1987 from Las Campanas, and 1991 and 1998 from CTIO. Measurin
g the shift of delicate Balmer filaments along the northwest rim of the remnant, we obtain proper motions of 280 +/- 8 mas/yr along the entire length where the filaments are well defined, with little systematic variation along the filaments. We also report very deep Halpha imaging observations of the entire remnant that clearly show very faint emission surrounding almost the entire shell, as well as some diffuse emission regions in the (projected) interior. Combining the proper motion measurement with a recent measurement of the shock velocity based on spectra of the same filaments by Ghavamian et al. leads to a distance of 2.17 +/- 0.08 kpc to SN1006. Several lines of argument suggest that SN1006 was a Type Ia event, so the improved distance measurement can be combined with the peak luminosity for SNeIa, as determined for events in galaxies with Cepheid-based distances, to calculate the apparent brightness of the spectacular event that drew wide attention in the eleventh century. The result, V_max = -7.5 =/- 0.4, lies squarely in the middle of the wide range of estimates based on the historical observations.
