Microsoft (Finland)

Similarity search is a primitive operation in database and Web search engines. With the advent of large-scale heterogeneous information networks that consist of multi-typed, interconnected objects, such as the bibliographic networks and social media networks, it is important to study similarity search in such networks. Intuitively, two objects are similar if they are linked by many paths in the network. However, most existing similarity measures are defined for homogeneous networks. Different semantic meanings behind paths are not taken into consideration. Thus they cannot be directly applied to heterogeneous networks. In this paper, we study similarity search that is defined among the same type of objects in heterogeneous networks. Moreover, by considering different linkage paths in a network, one could derive various similarity semantics. Therefore, we introduce the concept of meta path-based similarity , where a meta path is a path consisting of a sequence of relations defined between different object types ( i.e. , structural paths at the meta level). No matter whether a user would like to explicitly specify a path combination given sufficient domain knowledge, or choose the best path by experimental trials, or simply provide training examples to learn it, meta path forms a common base for a network-based similarity search engine. In particular, under the meta path framework we define a novel similarity measure called PathSim that is able to find peer objects in the network ( e.g. , find authors in the similar field and with similar reputation), which turns out to be more meaningful in many scenarios compared with random-walk based similarity measures. In order to support fast online query processing for PathSim queries, we develop an efficient solution that partially materializes short meta paths and then concatenates them online to compute top- k results. Experiments on real data sets demonstrate the effectiveness and efficiency of our proposed paradigm.

Recently a great deal of attention has focused on quantum computation following a sequence of results suggesting that quantum computers are more powerful than classical probabilistic computers. Following Shor's result that factoring and the extraction of discrete logarithms are both solvable in quantum polynomial time, it is natural to ask whether all of NP can be efficiently solved in quantum polynomial time. In this paper, we address this question by proving that relative to an oracle chosen uniformly at random, with probability 1, the class NP cannot be solved on a quantum Turing machine in time $o(2^{n/2})$. We also show that relative to a permutation oracle chosen uniformly at random, with probability 1, the class $NP \cap coNP$ cannot be solved on a quantum Turing machine in time $o(2^{n/3})$. The former bound is tight since recent work of Grover shows how to accept the class NP relative to any oracle on a quantum computer in time $O(2^{n/2})$.

Digital photography has made it possible to quickly and easily take a pair of images of low-light environments: one with flash to capture detail and one without flash to capture ambient illumination. We present a variety of applications that analyze and combine the strengths of such flash/no-flash image pairs. Our applications include denoising and detail transfer (to merge the ambient qualities of the no-flash image with the high-frequency flash detail), white-balancing (to change the color tone of the ambient image), continuous flash (to interactively adjust flash intensity), and red-eye removal (to repair artifacts in the flash image). We demonstrate how these applications can synthesize new images that are of higher quality than either of the originals.

not specify an Internet standard of any kind. Distribution of this memo is unlimited. Copyright Notice

ANSI SQL-92 [MS, ANSI] defines Isolation Levels in terms of phenomena: Dirty Reads, Non-Repeatable Reads, and Phantoms. This paper shows that these phenomena and the ANSI SQL definitions fail to properly characterize several popular isolation levels, including the standard locking implementations of the levels covered. Ambiguity in the statement of the phenomena is investigated and a more formal statement is arrived at; in addition new phenomena that better characterize isolation types are introduced. Finally, an important multiversion isolation type, called Snapshot Isolation, is defined.

This paper presents a detailed study of improving visual representations for vision language (VL) tasks and develops an improved object detection model to provide object-centric representations of images. Compared to the most widely used bottom-up and top-down model [2], the new model is bigger, better-designed for VL tasks, and pre-trained on much larger training corpora that combine multiple public annotated object detection datasets. Therefore, it can generate representations of a richer collection of visual objects and concepts. While previous VL research focuses mainly on improving the vision-language fusion model and leaves the object detection model improvement untouched, we show that visual features matter significantly in VL models. In our experiments we feed the visual features generated by the new object detection model into a Transformer-based VL fusion model OSCAR [20], and utilize an improved approach OSCAR+ to pre-train the VL model and fine-tune it on a wide range of downstream VL tasks. Our results show that the new visual features significantly improve the performance across all VL tasks, creating new state-of-the-art results on seven public benchmarks. Code, models and pre-extracted features are released at https://github.com/pzzhang/VinVL.

Companies providing cloud-scale services have an increasing need to store and analyze massive data sets such as search logs and click streams. For cost and performance reasons, processing is typically done on large clusters of shared-nothing commodity machines. It is imperative to develop a programming model that hides the complexity of the underlying system but provides flexibility by allowing users to extend functionality to meet a variety of requirements. In this paper, we present a new declarative and extensible scripting language, SCOPE (Structured Computations Optimized for Parallel Execution), targeted for this type of massive data analysis. The language is designed for ease of use with no explicit parallelism, while being amenable to efficient parallel execution on large clusters. SCOPE borrows several features from SQL. Data is modeled as sets of rows composed of typed columns. The select statement is retained with inner joins, outer joins, and aggregation allowed. Users can easily define their own functions and implement their own versions of operators: extractors (parsing and constructing rows from a file), processors (row-wise processing), reducers (group-wise processing), and combiners (combining rows from two inputs). SCOPE supports nesting of expressions but also allows a computation to be specified as a series of steps, in a manner often preferred by programmers. We also describe how scripts are compiled into efficient, parallel execution plans and executed on large clusters.

The purpose of this article is to present a set of best practices for competency modeling based on the experiences and lessons learned from the major perspectives on this topic (including applied, academic, and professional). Competency models are defined, and their key advantages are explained. Then, the many uses of competency models are described. The bulk of the article is a set of 20 best practices divided into 3 areas: analyzing competency information, organizing and presenting competency information, and using competency information. The best practices are described and explained, practice advice is provided, and then the best practices are illustrated with numerous practical examples. Finally, how competency modeling differs from and complements job analysis is explained throughout.

Most programs today are written not by professional software developers, but by people with expertise in other domains working towards goals for which they need computational support. For example, a teacher might write a grading spreadsheet to save time grading, or an interaction designer might use an interface builder to test some user interface design ideas. Although these end-user programmers may not have the same goals as professional developers, they do face many of the same software engineering challenges, including understanding their requirements, as well as making decisions about design, reuse, integration, testing, and debugging. This article summarizes and classifies research on these activities, defining the area of End-User Software Engineering (EUSE) and related terminology. The article then discusses empirical research about end-user software engineering activities and the technologies designed to support them. The article also addresses several crosscutting issues in the design of EUSE tools, including the roles of risk, reward, and domain complexity, and self-efficacy in the design of EUSE tools and the potential of educating users about software engineering principles.

We present the first end-to-end solution to create high-quality free-viewpoint video encoded as a compact data stream. Our system records performances using a dense set of RGB and IR video cameras, generates dynamic textured surfaces, and compresses these to a streamable 3D video format. Four technical advances contribute to high fidelity and robustness: multimodal multi-view stereo fusing RGB, IR, and silhouette information; adaptive meshing guided by automatic detection of perceptually salient areas; mesh tracking to create temporally coherent subsequences; and encoding of tracked textured meshes as an MPEG video stream. Quantitative experiments demonstrate geometric accuracy, texture fidelity, and encoding efficiency. We release several datasets with calibrated inputs and processed results to foster future research.

In previous papers [SC05, SBC+07], some of us predicted the end of size fits all as a commercial relational DBMS paradigm. These papers presented reasons and experimental evidence that showed that the major RDBMS vendors can be outperformed by 1--2 orders of magnitude by specialized engines in the data warehouse, stream processing, text, and scientific database markets. Assuming that specialized engines dominate these markets over time, the current relational DBMS code lines will be left with the business data processing (OLTP) market and hybrid markets where more than one kind of capability is required. In this paper we show that current RDBMSs can be beaten by nearly two orders of magnitude in the OLTP market as well. The experimental evidence comes comparing a new OLTP prototype, H-Store, which we have built at M.I.T. to a popular RDBMS on the standard transactional benchmark, TPC-C. We conclude that the current RDBMS code lines, while attempting to be a size fits all solution, in fact, excel at nothing. Hence, they are 25 year old legacy code lines that should be retired in favor of a collection of from scratch specialized engines. The DBMS vendors (and the research community) should start with a clean sheet of paper and design systems for tomorrow's requirements, not continue to push code lines and architectures designed for yesterday's needs.

Fair queuing is a technique that allows each flow passing through a network device to have fair share of network resources. previous schemes for fair queuing that achieved nearly perfect fairness were expensive to implement: specifically, the work required to process a packet in these schemes was O(log(n)), where n is the number of active flows. This is expensive at high speeds. On the other hand, cheaper approximations of fair queuing that have been reported in the literature exhibit unfair behavior. In this paper, we describe a new approximation of fair queuing, that we call Deficit Round Robin. Our scheme achieves nearly perfect fairness in terms of throughput, requires only O(1) work to process a packet, and is simple enough to implement in hardware. Deficit Round Robin is also applicable to other scheduling problems where servicing cannot be broken up into smaller units.

A big convergence of language, vision, and multimodal pretraining is emerging. In this work, we introduce a general-purpose multimodal foundation model BEIT-3, which achieves excellent transfer performance on both vision and vision-language tasks. Specifically, we advance the big convergence from three aspects: backbone architecture, pretraining task, and model scaling up. We use Multiway Transformers for general-purpose modeling, where the modular architecture enables both deep fusion and modality-specific encoding. Based on the shared backbone, we perform masked “language” modeling on images (Imglish), texts (English), and image-text pairs (“parallel sentences”) in a unified manner. Experimental results show that BEIT-3 obtains remarkable performance on object detection (COCO), semantic segmentation (ADE20K), image classification (ImageNet), visual reasoning (NLVR2), visual question answering (VQAv2), image captioning (COCO), and cross-modal retrieval (Flickr30K, COCO).

We cast the ranking problem as (1) multiple classification (“Mc”) (2) multiple or-dinal classification, which lead to computationally tractable learning algorithms for relevance ranking in Web search. We consider the DCG criterion (discounted cumulative gain), a standard quality measure in information retrieval. Our ap-proach is motivated by the fact that perfect classifications result in perfect DCG scores and the DCG errors are bounded by classification errors. We propose us-ing the Expected Relevance to convert class probabilities into ranking scores. The class probabilities are learned using a gradient boosting tree algorithm. Evalua-tions on large-scale datasets show that our approach can improve LambdaRank [5] and the regressions-based ranker [6], in terms of the (normalized) DCG scores. An efficient implementation of the boosting tree algorithm is also presented. 1

Hyperscale datacenter providers have struggled to balance the growing need for specialized hardware (efficiency) with the economic benefits of homogeneity (manageability). In this paper we propose a new cloud architecture that uses reconfigurable logic to accelerate both network plane functions and applications. This Configurable Cloud architecture places a layer of reconfigurable logic (FPGAs) between the network switches and the servers, enabling network flows to be programmably transformed at line rate, enabling acceleration of local applications running on the server, and enabling the FPGAs to communicate directly, at datacenter scale, to harvest remote FPGAs unused by their local servers. We deployed this design over a production server bed, and show how it can be used for both service acceleration (Web search ranking) and network acceleration (encryption of data in transit at high-speeds). This architecture is much more scalable than prior work which used secondary rack-scale networks for inter-FPGA communication. By coupling to the network plane, direct FPGA-to-FPGA messages can be achieved at comparable latency to previous work, without the secondary network. Additionally, the scale of direct inter-FPGA messaging is much larger. The average round-trip latencies observed in our measurements among 24, 1000, and 250,000 machines are under 3, 9, and 20 microseconds, respectively. The Configurable Cloud architecture has been deployed at hyperscale in Microsoft's production datacenters worldwide.

Digital Health in the 21st Century Over the past several decades, the development and accelerated advancement of digital technology has prompted change across virtually all aspects of human endeavor. […]

Dynamic inference techniques have been demonstrated to provide useful support for various software engineering tasks including bug finding, test suite evaluation and improvement, and specification generation. To date, however, dynamic inference has only been used effectively on small programs under controlled conditions. In this paper, we identify reasons why scaling dynamic inference techniques has proven difficult, and introduce solutions that enable a dynamic inference technique to scale to large programs and work effectively with the imperfect traces typically available in industrial scenarios. We describe our approximate inference algorithm, present and evaluate heuristics for winnowing the large number of inferred properties to a manageable set of interesting properties, and report on experiments using inferred properties. We evaluate our techniques on JBoss and the Windows kernel. Our tool is able to infer many of the properties checked by the Static Driver Verifier and leads us to discover a previously unknown bug in Windows.

We introduce a hierarchical labeling scheme called ORDPATH that is implemented in the upcoming version of Microsoft® SQL Server™. ORDPATH labels nodes of an XML tree without requiring a schema (the most general case---a schema simplifies the problem). An example of an ORDPATH value display format is "1.5.3.9.1". A compressed binary representation of ORDPATH provides document order by simple byte-by-byte comparison and ancestry relationship equally simply. In addition, the ORDPATH scheme supports insertion of new nodes at arbitrary positions in the XML tree, their ORDPATH values "careted in" between ORDPATHs of sibling nodes, without relabeling any old nodes.

Benchmarking is an essential aspect of any database management system (DBMS) effort. Despite several recent advancements, such as pre-configured cloud database images and database-as-a-service (DBaaS) offerings, the deployment of a comprehensive testing platform with a diverse set of datasets and workloads is still far from being trivial. In many cases, researchers and developers are limited to a small number of workloads to evaluate the performance characteristics of their work. This is due to the lack of a universal benchmarking infrastructure, and to the difficulty of gaining access to real data and workloads. This results in lots of unnecessary engineering efforts and makes the performance evaluation results difficult to compare. To remedy these problems, we present OLTP-Bench, an extensible "batteries included" DBMS benchmarking testbed. The key contributions of OLTP-Bench are its ease of use and extensibility, support for tight control of transaction mixtures, request rates, and access distributions over time, as well as the ability to support all major DBMSs and DBaaS platforms. Moreover, it is bundled with fifteen workloads that all differ in complexity and system demands, including four synthetic workloads, eight workloads from popular benchmarks, and three workloads that are derived from real-world applications. We demonstrate through a comprehensive set of experiments conducted on popular DBMS and DBaaS offerings the different features provided by OLTP-Bench and the effectiveness of our testbed in characterizing the performance of database services.

Bugs in kernel-level device drivers cause 85% of the system crashes in the Windows XP operating system [44]. One of the sources of these errors is the complexity of the Windows driver API itself: programmers must master a complex set of rules about how to use the driver API in order to create drivers that are good clients of the kernel. We have built a static analysis engine that finds API usage errors in C programs. The Static Driver Verifier tool (SDV) uses this engine to find kernel API usage errors in a driver. SDV includes models of the OS and the environment of the device driver, and over sixty API usage rules. SDV is intended to be used by driver developers "out of the box." Thus, it has stringent requirements: (1) complete automation with no input from the user; (2) a low rate of false errors. We discuss the techniques used in SDV to meet these requirements, and empirical results from running SDV on over one hundred Windows device drivers.