Unlike the conventional wisdom in statistical learning theory, the test error of a deep neural network (DNN) often demonstrates double descent: as the model complexity increases, it first follows a classical U-shaped curve and then shows a second descent. Through bias-variance decomposition, recent studies revealed that the bell-shaped variance is the major cause of model-wise double descent (when the DNN is widened gradually). This paper investigates epoch-wise double descent, i.e., the test error of a DNN also shows double descent as the number of training epoches increases. By extending the bias-variance analysis to epoch-wise double descent of the zero-one loss, we surprisingly find that the variance itself, without the bias, varies consistently with the test error. Inspired by this result, we propose a novel metric, optimization variance (OV), to measure the diversity of model updates caused by the stochastic gradients of random training batches drawn in the same iteration. OV can be estimated using samples from the training set only but correlates well with the (unknown) \emph{test} error, and hence early stopping may be achieved without using a validation set.
In modern internet industries, deep learning based recommender systems have became an indispensable building block for a wide spectrum of applications, such as search engine, news feed, and short video clips. However, it remains challenging to carry the well-trained deep models for online real-time inference serving, with respect to the time-varying web-scale traffics from billions of users, in a cost-effective manner. In this work, we present JIZHI - a Model-as-a-Service system - that per second handles hundreds of millions of online inference requests to huge deep models with more than trillions of sparse parameters, for over twenty real-time recommendation services at Baidu, Inc. In JIZHI, the inference workflow of every recommendation request is transformed to a Staged Event-Driven Pipeline (SEDP), where each node in the pipeline refers to a staged computation or I/O intensive task processor. With traffics of real-time inference requests arrived, each modularized processor can be run in a fully asynchronized way and managed separately. Besides, JIZHI introduces heterogeneous and hierarchical storage to further accelerate the online inference process by reducing unnecessary computations and potential data access latency induced by ultra-sparse model parameters. Moreover, an intelligent resource manager has been deployed to maximize the throughput of JIZHI over the shared infrastructure by searching the optimal resource allocation plan from historical logs and fine-tuning the load shedding policies over intermediate system feedback. Extensive experiments have been done to demonstrate the advantages of JIZHI from the perspectives of end-to-end service latency, system-wide throughput, and resource consumption. JIZHI has helped Baidu saved more than ten million US dollars in hardware and utility costs while handling 200% more traffics without sacrificing inference efficiency.
In recent years, data and computing resources are typically distributed in the devices of end users, various regions or organizations. Because of laws or regulations, the distributed data and computing resources cannot be directly shared among different regions or organizations for machine learning tasks. Federated learning emerges as an efficient approach to exploit distributed data and computing resources, so as to collaboratively train machine learning models, while obeying the laws and regulations and ensuring data security and data privacy. In this paper, we provide a comprehensive survey of existing works for federated learning. We propose a functional architecture of federated learning systems and a taxonomy of related techniques. Furthermore, we present the distributed training, data communication, and security of FL systems. Finally, we analyze their limitations and propose future research directions.
To improve the performance of deep learning, mixup has been proposed to force the neural networks favoring simple linear behaviors in-between training samples. Performing mixup for transfer learning with pre-trained models however is not that simple, a high capacity pre-trained model with a large fully-connected (FC) layer could easily overfit to the target dataset even with samples-to-labels mixed up. In this work, we propose SMILE - Self-Distilled Mixup for EffIcient Transfer LEarning. With mixed images as inputs, SMILE regularizes the outputs of CNN feature extractors to learn from the mixed feature vectors of inputs (sample-to-feature mixup), in addition to the mixed labels. Specifically, SMILE incorporates a mean teacher, inherited from the pre-trained model, to provide the feature vectors of input samples in a self-distilling fashion, and mixes up the feature vectors accordingly via a novel triplet regularizer. The triple regularizer balances the mixup effects in both feature and label spaces while bounding the linearity in-between samples for pre-training tasks. Extensive experiments have been done to verify the performance improvement made by SMILE, in comparisons with a wide spectrum of transfer learning algorithms, including fine-tuning, L2-SP, DELTA, and RIFLE, even with mixup strategies combined. Ablation studies show that the vanilla sample-to-label mixup strategies could marginally increase the linearity in-between training samples but lack of generalizability, while SMILE significantly improve the mixup effects in both label and feature spaces with both training and testing datasets. The empirical observations backup our design intuition and purposes.
Deep neural networks have been well-known for their superb performance in handling various machine learning and artificial intelligence tasks. However, due to their over-parameterized black-box nature, it is often difficult to understand the prediction results of deep models. In recent years, many interpretation tools have been proposed to explain or reveal the ways that deep models make decisions. In this paper, we review this line of research and try to make a comprehensive survey. Specifically, we introduce and clarify two basic concepts-interpretations and interpretability-that people usually get confused. First of all, to address the research efforts in interpretations, we elaborate the design of several recent interpretation algorithms, from different perspectives, through proposing a new taxonomy. Then, to understand the results of interpretation, we also survey the performance metrics for evaluating interpretation algorithms. Further, we summarize the existing work in evaluating models' interpretability using "trustworthy" interpretation algorithms. Finally, we review and discuss the connections between deep models' interpretations and other factors, such as adversarial robustness and data augmentations, and we introduce several open-source libraries for interpretation algorithms and evaluation approaches.
The novel coronavirus disease (COVID-19) has crushed daily routines and is still rampaging through the world. Existing solution for nonpharmaceutical interventions usually needs to timely and precisely select a subset of residential urban areas for containment or even quarantine, where the spatial distribution of confirmed cases has been considered as a key criterion for the subset selection. While such containment measure has successfully stopped or slowed down the spread of COVID-19 in some countries, it is criticized for being inefficient or ineffective, as the statistics of confirmed cases are usually time-delayed and coarse-grained. To tackle the issues, we propose C-Watcher, a novel data-driven framework that aims at screening every neighborhood in a target city and predicting infection risks, prior to the spread of COVID-19 from epicenters to the city. In terms of design, C-Watcher collects large-scale long-term human mobility data from Baidu Maps, then characterizes every residential neighborhood in the city using a set of features based on urban mobility patterns. Furthermore, to transfer the firsthand knowledge (witted in epicenters) to the target city before local outbreaks, we adopt a novel adversarial encoder framework to learn "city-invariant" representations from the mobility-related features for precise early detection of high-risk neighborhoods, even before any confirmed cases known, in the target city. We carried out extensive experiments on C-Watcher using the real-data records in the early stage of COVID-19 outbreaks, where the results demonstrate the efficiency and effectiveness of C-Watcher for early detection of high-risk neighborhoods from a large number of cities.
Accurately predicting the binding affinity between drugs and proteins is an essential step for computational drug discovery. Since graph neural networks (GNNs) have demonstrated remarkable success in various graph-related tasks, GNNs have been considered as a promising tool to improve the binding affinity prediction in recent years. However, most of the existing GNN architectures can only encode the topological graph structure of drugs and proteins without considering the relative spatial information among their atoms. Whereas, different from other graph datasets such as social networks and commonsense knowledge graphs, the relative spatial position and chemical bonds among atoms have significant impacts on the binding affinity. To this end, in this paper, we propose a diStance-aware Molecule graph Attention Network (S-MAN) tailored to drug-target binding affinity prediction. As a dedicated solution, we first propose a position encoding mechanism to integrate the topological structure and spatial position information into the constructed pocket-ligand graph. Moreover, we propose a novel edge-node hierarchical attentive aggregation structure which has edge-level aggregation and node-level aggregation. The hierarchical attentive aggregation can capture spatial dependencies among atoms, as well as fuse the position-enhanced information with the capability of discriminating multiple spatial relations among atoms. Finally, we conduct extensive experiments on two standard datasets to demonstrate the effectiveness of S-MAN.
Fine-tuning deep neural networks pre-trained on large scale datasets is one of the most practical transfer learning paradigm given limited quantity of training samples. To obtain better generalization, using the starting point as the reference, either through weights or features, has been successfully applied to transfer learning as a regularizer. However, due to the domain discrepancy between the source and target tasks, there exists obvious risk of negative transfer. In this paper, we propose a novel transfer learning algorithm, introducing the idea of Target-awareness REpresentation Disentanglement (TRED), where the relevant knowledge with respect to the target task is disentangled from the original source model and used as a regularizer during fine-tuning the target model. Experiments on various real world datasets show that our method stably improves the standard fine-tuning by more than 2% in average. TRED also outperforms other state-of-the-art transfer learning regularizers such as L2-SP, AT, DELTA and BSS.