Current PB solvers implement many techniques inspired by the CDCL architecture of modern SAT solvers, so as to benefit from its practical efficiency. However, they also need to deal with the fact that many of the properties leveraged by this architecture are no longer true when considering PB constraints. In this paper, we focus on one of these properties, namely the optimality of the so-called first unique implication point (1-UIP). While it is well known that learning the first assertive clause produced during conflict analysis ensures to perform the highest possible backjump in a SAT solver, we show that there is no such guarantee in the presence of PB constraints. We also introduce and evaluate different approaches designed to improve the backjump level identified during conflict analysis by allowing to continue the analysis after reaching the 1-UIP. Our experiments show that sub-optimal backjumps are fairly common in PB solvers, even though their impact on the solver is not clear.
相關內容
When to initiate treatment on patients is an important problem in many medical studies such as AIDS and cancer. In this article, we formulate the treatment initiation time problem for time-to-event data and propose an optimal individualized regime that determines the best treatment initiation time for individual patients based on their characteristics. Different from existing optimal treatment regimes where treatments are undertaken at a pre-specified time, here new challenges arise from the complicated missing mechanisms in treatment initiation time data and the continuous treatment rule in terms of initiation time. To tackle these challenges, we propose to use restricted mean residual lifetime as a value function to evaluate the performance of different treatment initiation regimes, and develop a nonparametric estimator for the value function, which is consistent even when treatment initiation times are not completely observable and their distribution is unknown. We also establish the asymptotic properties of the resulting estimator in the decision rule and its associated value function estimator. In particular, the asymptotic distribution of the estimated value function is nonstandard, which follows a weighted chi-squared distribution. The finite-sample performance of the proposed method is evaluated by simulation studies and is further illustrated with an application to a breast cancer data.
Due to its role in many computer vision tasks, image matching has been subjected to an active investigation by researchers, which has lead to better and more discriminant feature descriptors and to more robust matching strategies, also thanks to the advent of the deep learning and the increased computational power of the modern hardware. Despite of these achievements, the keypoint extraction process at the base of the image matching pipeline has not seen equivalent progresses. This paper presents Harrisz$^{+}$, an upgrade to the HarrisZ corner detector, optimized to synergically take advance of the recent improvements of the other steps of the image matching pipeline. Harrisz$^{+}$ does not only consists of a tuning of the setup parameters, but introduces further refinements to the selection criteria delineated by HarrisZ, so providing more, yet discriminative, keypoints, which are better distributed on the image and with higher localization accuracy. The image matching pipeline including Harrisz$^{+}$, together with the other modern components, obtained in different recent matching benchmarks state-of-the-art results among the classic image matching pipelines, closely following results of the more recent fully deep end-to-end trainable approaches.
The successful training of neural networks typically involves careful and time consuming hyperparameter tuning. Population Based Training (PBT) has recently been proposed to automate this process. PBT trains a population of neural networks concurrently, frequently mutating their hyperparameters throughout their training. However, the decision mechanisms of PBT are greedy and favour short-term improvements which can, in some cases, lead to poor long-term performance. This paper presents Faster Improvement Rate PBT (FIRE PBT) which addresses this problem. Our method is guided by an assumption: given two neural networks with similar performance and training with similar hyperparameters, the network showing the faster rate of improvement will lead to a better final performance. Using this, we derive a novel fitness metric and use it to make some of the population members focus on long-term performance. Our experiments show that FIRE PBT is able to outperform PBT on the ImageNet benchmark and match the performance of networks that were trained with a hand-tuned learning rate schedule. We apply FIRE PBT to reinforcement learning tasks and show that it leads to faster learning and higher final performance than both PBT and random hyperparameter search.
Monte Carlo (MC) methods are important computational tools for molecular structure optimizations and predictions. When solvent effects are explicitly considered, MC methods become very expensive due to the large degree of freedom associated with the water molecules and mobile ions. Alternatively implicit-solvent MC can largely reduce the computational cost by applying a mean field approximation to solvent effects and meanwhile maintains the atomic detail of the target molecule. The two most popular implicit-solvent models are the Poisson-Boltzmann (PB) model and the Generalized Born (GB) model in a way such that the GB model is an approximation to the PB model but is much faster in simulation time. In this work, we develop a machine learning-based implicit-solvent Monte Carlo (MLIMC) method by combining the advantages of both implicit solvent models in accuracy and efficiency. Specifically, the MLIMC method uses a fast and accurate PB-based machine learning (PBML) scheme to compute the electrostatic solvation free energy at each step. We validate our MLIMC method by using a benzene-water system and a protein-water system. We show that the proposed MLIMC method has great advantages in speed and accuracy for molecular structure optimization and prediction.
Analysts are often confronted with censoring, wherein some variables are not observed at their true value, but rather at a value that is known to fall above or below that truth. While much attention has been given to the analysis of censored outcomes, contemporary focus has shifted to censored covariates, as well. Missing data is often overcome using multiple imputation, which leverages the entire dataset by replacing missing values with informed placeholders, and this method can be modified for censored data by also incorporating partial information from censored values. One such modification involves replacing censored covariates with their conditional means given other fully observed information, such as the censored value or additional covariates. So-called conditional mean imputation approaches were proposed for censored covariates in Atem et al. [2017], Atem et al.[2019a], and Atem et al. [2019b]. These methods are robust to additional parametric assumptions on the censored covariate and utilize all available data, which is appealing. As we worked to implement these methods, however, we discovered that these three manuscripts provide nonequivalent formulas and, in fact, none is the correct formula for the conditional mean. Herein, we derive the correct form of the conditional mean and demonstrate the impact of the incorrect formulas on the imputed values and statistical inference. Under several settings considered, using an incorrect formula is seen to seriously bias parameter estimation in simple linear regression. Lastly, we provide user-friendly R software, the imputeCensoRd package, to enable future researchers to tackle censored covariates in their data.
A core capability of intelligent systems is the ability to quickly learn new tasks by drawing on prior experience. Gradient (or optimization) based meta-learning has recently emerged as an effective approach for few-shot learning. In this formulation, meta-parameters are learned in the outer loop, while task-specific models are learned in the inner-loop, by using only a small amount of data from the current task. A key challenge in scaling these approaches is the need to differentiate through the inner loop learning process, which can impose considerable computational and memory burdens. By drawing upon implicit differentiation, we develop the implicit MAML algorithm, which depends only on the solution to the inner level optimization and not the path taken by the inner loop optimizer. This effectively decouples the meta-gradient computation from the choice of inner loop optimizer. As a result, our approach is agnostic to the choice of inner loop optimizer and can gracefully handle many gradient steps without vanishing gradients or memory constraints. Theoretically, we prove that implicit MAML can compute accurate meta-gradients with a memory footprint that is, up to small constant factors, no more than that which is required to compute a single inner loop gradient and at no overall increase in the total computational cost. Experimentally, we show that these benefits of implicit MAML translate into empirical gains on few-shot image recognition benchmarks.
Although the Transformer translation model (Vaswani et al., 2017) has achieved state-of-the-art performance in a variety of translation tasks, how to use document-level context to deal with discourse phenomena problematic for Transformer still remains a challenge. In this work, we extend the Transformer model with a new context encoder to represent document-level context, which is then incorporated into the original encoder and decoder. As large-scale document-level parallel corpora are usually not available, we introduce a two-step training method to take full advantage of abundant sentence-level parallel corpora and limited document-level parallel corpora. Experiments on the NIST Chinese-English datasets and the IWSLT French-English datasets show that our approach improves over Transformer significantly.
Homographs, words with different meanings but the same surface form, have long caused difficulty for machine translation systems, as it is difficult to select the correct translation based on the context. However, with the advent of neural machine translation (NMT) systems, which can theoretically take into account global sentential context, one may hypothesize that this problem has been alleviated. In this paper, we first provide empirical evidence that existing NMT systems in fact still have significant problems in properly translating ambiguous words. We then proceed to describe methods, inspired by the word sense disambiguation literature, that model the context of the input word with context-aware word embeddings that help to differentiate the word sense be- fore feeding it into the encoder. Experiments on three language pairs demonstrate that such models improve the performance of NMT systems both in terms of BLEU score and in the accuracy of translating homographs.
Latent Dirichlet Allocation(LDA) is a popular topic model. Given the fact that the input corpus of LDA algorithms consists of millions to billions of tokens, the LDA training process is very time-consuming, which may prevent the usage of LDA in many scenarios, e.g., online service. GPUs have benefited modern machine learning algorithms and big data analysis as they can provide high memory bandwidth and computation power. Therefore, many frameworks, e.g. Ten- sorFlow, Caffe, CNTK, support to use GPUs for accelerating the popular machine learning data-intensive algorithms. However, we observe that LDA solutions on GPUs are not satisfying. In this paper, we present CuLDA_CGS, a GPU-based efficient and scalable approach to accelerate large-scale LDA problems. CuLDA_CGS is designed to efficiently solve LDA problems at high throughput. To it, we first delicately design workload partition and synchronization mechanism to exploit the benefits of mul- tiple GPUs. Then, we offload the LDA sampling process to each individual GPU by optimizing from the sampling algorithm, par- allelization, and data compression perspectives. Evaluations show that compared with state-of-the-art LDA solutions, CuLDA_CGS outperforms them by a large margin (up to 7.3X) on a single GPU. CuLDA_CGS is able to achieve extra 3.0X speedup on 4 GPUs. The source code is publicly available on //github.com/cuMF/ CuLDA_CGS.
Current image captioning methods are usually trained via (penalized) maximum likelihood estimation. However, the log-likelihood score of a caption does not correlate well with human assessments of quality. Standard syntactic evaluation metrics, such as BLEU, METEOR and ROUGE, are also not well correlated. The newer SPICE and CIDEr metrics are better correlated, but have traditionally been hard to optimize for. In this paper, we show how to use a policy gradient (PG) method to directly optimize a linear combination of SPICE and CIDEr (a combination we call SPIDEr): the SPICE score ensures our captions are semantically faithful to the image, while CIDEr score ensures our captions are syntactically fluent. The PG method we propose improves on the prior MIXER approach, by using Monte Carlo rollouts instead of mixing MLE training with PG. We show empirically that our algorithm leads to easier optimization and improved results compared to MIXER. Finally, we show that using our PG method we can optimize any of the metrics, including the proposed SPIDEr metric which results in image captions that are strongly preferred by human raters compared to captions generated by the same model but trained to optimize MLE or the COCO metrics.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191