Multiplying matrices is among the most fundamental and compute-intensive operations in machine learning. Consequently, there has been significant work on efficiently approximating matrix multiplies. We introduce a learning-based algorithm for this task that greatly outperforms existing methods. Experiments using hundreds of matrices from diverse domains show that it often runs $100\times$ faster than exact matrix products and $10\times$ faster than current approximate methods. In the common case that one matrix is known ahead of time, our method also has the interesting property that it requires zero multiply-adds. These results suggest that a mixture of hashing, averaging, and byte shuffling$-$the core operations of our method$-$could be a more promising building block for machine learning than the sparsified, factorized, and/or scalar quantized matrix products that have recently been the focus of substantial research and hardware investment.
Nonlocal models, including peridynamics, often use integral operators that embed lengthscales in their definition. However, the integrands in these operators are difficult to define from the data that are typically available for a given physical system, such as laboratory mechanical property tests. In contrast, molecular dynamics (MD) does not require these integrands, but it suffers from computational limitations in the length and time scales it can address. To combine the strengths of both methods and to obtain a coarse-grained, homogenized continuum model that efficiently and accurately captures materials' behavior, we propose a learning framework to extract, from MD data, an optimal Linear Peridynamic Solid (LPS) model as a surrogate for MD displacements. To maximize the accuracy of the learnt model we allow the peridynamic influence function to be partially negative, while preserving the well-posedness of the resulting model. To achieve this, we provide sufficient well-posedness conditions for discretized LPS models with sign-changing influence functions and develop a constrained optimization algorithm that minimizes the equation residual while enforcing such solvability conditions. This framework guarantees that the resulting model is mathematically well-posed, physically consistent, and that it generalizes well to settings that are different from the ones used during training. We illustrate the efficacy of the proposed approach with several numerical tests for single layer graphene. Our two-dimensional tests show the robustness of the proposed algorithm on validation data sets that include thermal noise, different domain shapes and external loadings, and discretizations substantially different from the ones used for training.
Reinforcement learning (RL) is an effective technique for training decision-making agents through interactions with their environment. The advent of deep learning has been associated with highly notable successes with sequential decision making problems - such as defeating some of the highest-ranked human players at Go. In digital advertising, real-time bidding (RTB) is a common method of allocating advertising inventory through real-time auctions. Bidding strategies need to incorporate logic for dynamically adjusting parameters in order to deliver pre-assigned campaign goals. Here we discuss techniques toward using RL to train bidding agents. As a campaign metric we particularly focused on viewability: the percentage of inventory which goes on to be viewed by an end user. This paper is presented as a survey of techniques and experiments which we developed through the course of this research. We discuss expanding our training data to include edge cases by training on simulated interactions. We discuss the experimental results comparing the performance of several promising RL algorithms, and an approach to hyperparameter optimization of an actor/critic training pipeline through Bayesian optimization. Finally, we present live-traffic tests of some of our RL agents against a rule-based feedback-control approach, demonstrating the potential for this method as well as areas for further improvement. This paper therefore presents an arrangement of our findings in this quickly developing field, and ways that it can be applied to an RTB use case.
As technology scaling is approaching the physical limit, lithography hotspot detection has become an essential task in design for manufacturability. While the deployment of pattern matching or machine learning in hotspot detection can help save significant simulation time, such methods typically demand for non-trivial quality data to build the model, which most design houses are short of. Moreover, the design houses are also unwilling to directly share such data with the other houses to build a unified model, which can be ineffective for the design house with unique design patterns due to data insufficiency. On the other hand, with data homogeneity in each design house, the locally trained models can be easily over-fitted, losing generalization ability and robustness. In this paper, we propose a heterogeneous federated learning framework for lithography hotspot detection that can address the aforementioned issues. On one hand, the framework can build a more robust centralized global sub-model through heterogeneous knowledge sharing while keeping local data private. On the other hand, the global sub-model can be combined with a local sub-model to better adapt to local data heterogeneity. The experimental results show that the proposed framework can overcome the challenge of non-independent and identically distributed (non-IID) data and heterogeneous communication to achieve very high performance in comparison to other state-of-the-art methods while guaranteeing a good convergence rate in various scenarios.
In the Chinese medical insurance industry, the assessor's role is essential and requires significant efforts to converse with the claimant. This is a highly professional job that involves many parts, such as identifying personal information, collecting related evidence, and making a final insurance report. Due to the coronavirus (COVID-19) pandemic, the previous offline insurance assessment has to be conducted online. However, for the junior assessor often lacking practical experience, it is not easy to quickly handle such a complex online procedure, yet this is important as the insurance company needs to decide how much compensation the claimant should receive based on the assessor's feedback. In order to promote assessors' work efficiency and speed up the overall procedure, in this paper, we propose a dialogue-based information extraction system that integrates advanced NLP technologies for medical insurance assessment. With the assistance of our system, the average time cost of the procedure is reduced from 55 minutes to 35 minutes, and the total human resources cost is saved 30% compared with the previous offline procedure. Until now, the system has already served thousands of online claim cases.
Guided depth super-resolution is a practical task where a low-resolution and noisy input depth map is restored to a high-resolution version, with the help of a high-resolution RGB guide image. Existing methods usually view this task as a generalized guided filtering problem that relies on designing explicit filters and objective functions, or a dense regression problem that directly predicts the target image via deep neural networks. These methods suffer from either model capability or interpretability. Inspired by the recent progress in implicit neural representation, we propose to formulate the guided super-resolution as a neural implicit image interpolation problem, where we take the form of a general image interpolation but use a novel Joint Implicit Image Function (JIIF) representation to learn both the interpolation weights and values. JIIF represents the target image domain with spatially distributed local latent codes extracted from the input image and the guide image, and uses a graph attention mechanism to learn the interpolation weights at the same time in one unified deep implicit function. We demonstrate the effectiveness of our JIIF representation on guided depth super-resolution task, significantly outperforming state-of-the-art methods on three public benchmarks. Code can be found at \url{https://git.io/JC2sU}.
Esports have become major international sports with hundreds of millions of spectators. Esports games generate massive amounts of telemetry data. Using these to predict the outcome of esports matches has received considerable attention, but micro-predictions, which seek to predict events inside a match, is as yet unknown territory. Micro-predictions are however of perennial interest across esports commentators and audience, because they provide the ability to observe events that might otherwise be missed: esports games are highly complex with fast-moving action where the balance of a game can change in the span of seconds, and where events can happen in multiple areas of the playing field at the same time. Such events can happen rapidly, and it is easy for commentators and viewers alike to miss an event and only observe the following impact of events. In Dota 2, a player hero being killed by the opposing team is a key event of interest to commentators and audience. We present a deep learning network with shared weights which provides accurate death predictions within a five-second window. The network is trained on a vast selection of Dota 2 gameplay features and professional/semi-professional level match dataset. Even though death events are rare within a game (1\% of the data), the model achieves 0.377 precision with 0.725 recall on test data when prompted to predict which of any of the 10 players of either team will die within 5 seconds. An example of the system applied to a Dota 2 match is presented. This model enables real-time micro-predictions of kills in Dota 2, one of the most played esports titles in the world, giving commentators and viewers time to move their attention to these key events.
Generative adversarial networks (GANs) have shown promising results when applied on partial differential equations and financial time series generation. We investigate if GANs can also be used to approximate one-dimensional Ito stochastic differential equations (SDEs). We propose a scheme that approximates the path-wise conditional distribution of SDEs for large time steps. Standard GANs are only able to approximate processes in distribution, yielding a weak approximation to the SDE. A conditional GAN architecture is proposed that enables strong approximation. We inform the discriminator of this GAN with the map between the prior input to the generator and the corresponding output samples, i.e. we introduce a `supervised GAN'. We compare the input-output map obtained with the standard GAN and supervised GAN and show experimentally that the standard GAN may fail to provide a path-wise approximation. The GAN is trained on a dataset obtained with exact simulation. The architecture was tested on geometric Brownian motion (GBM) and the Cox-Ingersoll-Ross (CIR) process. The supervised GAN outperformed the Euler and Milstein schemes in strong error on a discretisation with large time steps. It also outperformed the standard conditional GAN when approximating the conditional distribution. We also demonstrate how standard GANs may give rise to non-parsimonious input-output maps that are sensitive to perturbations, which motivates the need for constraints and regularisation on GAN generators.
In applications such as object tracking, time-series data inevitably carry missing observations. Following the success of deep learning-based models for various sequence learning tasks, these models increasingly replace classic approaches in object tracking applications for inferring the objects' motion states. While traditional tracking approaches can deal with missing observations, most of their deep counterparts are, by default, not suited for this. Towards this end, this paper introduces a transformer-based approach for handling missing observations in variable input length trajectory data. The model is formed indirectly by successively increasing the complexity of the demanded inference tasks. Starting from reproducing noise-free trajectories, the model then learns to infer trajectories from noisy inputs. By providing missing tokens, binary-encoded missing events, the model learns to in-attend to missing data and infers a complete trajectory conditioned on the remaining inputs. In the case of a sequence of successive missing events, the model then acts as a pure prediction model. The abilities of the approach are demonstrated on synthetic data and real-world data reflecting prototypical object tracking scenarios.
Complex multivariate time series arise in many fields, ranging from computer vision to robotics or medicine. Often we are interested in the independent underlying factors that give rise to the high-dimensional data we are observing. While many models have been introduced to learn such disentangled representations, only few attempt to explicitly exploit the structure of sequential data. We investigate the disentanglement properties of Gaussian process variational autoencoders, a class of models recently introduced that have been successful in different tasks on time series data. Our model exploits the temporal structure of the data by modeling each latent channel with a GP prior and employing a structured variational distribution that can capture dependencies in time. We demonstrate the competitiveness of our approach against state-of-the-art unsupervised and weakly-supervised disentanglement methods on a benchmark task. Moreover, we provide evidence that we can learn meaningful disentangled representations on real-world medical time series data.