Large-scale optimization has become an important application for data management systems, particularly in the context of statistical machine learning. In this paper, we consider how one might implement the join-and-co-group pattern in the context of a fully declarative data processing system. The join-and-co-group pattern is ubiquitous in iterative, large-scale optimization. In the join-and-co-group pattern, a user-defined function g is parameterized with a data object x as well as the subset of the statistical model Θxthat applies to that object, so that g(x|Θx) can be used to compute a partial update of the model. This is repeated for every x in the full data set X. All partial updates are then aggregated and used to perform a complete update of the model. The join-and-co-group pattern has several implementation challenges, including the potential for a massive blow-up in the size of a fully parameterized model. Thus, unless the correct physical execution plan be chosen for implementing the join-and-co-group pattern, it is easily possible to have an execution that takes a very long time or even fails to complete. In this paper, we carefully consider the alternatives for implementing the join-and-co-group pattern on top of a declarative system, as well as how the best alternative can be selected automatically. Our focus is on the SimSQL database system, which is an SQL-based system with special facilities for large-scale, iterative optimization. Since it is an SQL-based system with a query optimizer, those choices can be made automatically.

Many machine learning (ML) systems allow the specification of ML algorithms by means of linear algebra programs, and automatically generate efficient execution plans. The opportunities for fused operators---in terms of fused chains of basic operators---are ubiquitous, and include fewer materialized intermediates, fewer scans of inputs, and sparsity exploitation across operators. However, existing fusion heuristics struggle to find good plans for complex operator DAGs or hybrid plans of local and distributed operations. In this paper, we introduce an exact yet practical cost-based optimization framework for fusion plans and describe its end-to-end integration into Apache SystemML. We present techniques for candidate exploration and selection of fusion plans, as well as code generation of local and distributed operations over dense, sparse, and compressed data. Our experiments in SystemML show end-to-end performance improvements of up to 22x, with negligible compilation overhead.

The rising need for custom machine learning (ML) algorithms and the growing data sizes that require the exploitation of distributed, data-parallel frameworks such as MapReduce or Spark, pose significant productivity challenges to data scientists. Apache SystemML addresses these challenges through declarative ML by (1) increasing the productivity of data scientists as they are able to express custom algorithms in a familiar domain-specific language covering linear algebra primitives and statistical functions, and (2) transparently running these ML algorithms on distributed, data-parallel frameworks by applying cost-based compilation techniques to generate efficient, low-level execution plans with in-memory single-node and large-scale distributed operations. This paper describes SystemML on Apache Spark, end to end, including insights into various optimizer and runtime techniques as well as performance characteristics. We also share lessons learned from porting SystemML to Spark and declarative ML in general. Finally, SystemML is open-source, which allows the database community to leverage it as a testbed for further research.

Virtually all work on topic modeling has assumed that the topics are to be learned over a text-based document corpus. However, there exist important applications where topic models must be learned over an audio corpus of spoken language. Unfortunately, speech-to-text programs can have very low accuracy. We therefore propose a novel topic model for spoken language that incorporates a statistical model of speech-to-text software behavior. Crucially, our model exploits the uncertainty numbers returned by the software. Our ideas apply to any domain in which it would be useful to build a topic model over data in which uncertainties are explicitly represented.

Spoken Web is an alternative Web for low-literacy users in the developing world. People can create audio content over phone and share on the Spoken Web. This enables easy creation of locally relevant content. Even on the World Wide Web in developed regions, the recent increase in traffic is due to the locally relevant content created on social networking sites. This paper argues that content search and ranking in the new scenario needs a re-look. The generic model of using in-links for ranking such content is not an appropriate measure of the content relevance in such a collaborative Web 2.0 world. This paper aims to bring the social context in Spoken Web ranking. We formulate a relationship function between the query-creator and the content-creator and use this as one measure of the content relevance to the user. The relationship function uses the geographical location of the two people and their prior browsing preferences as parameters to determine the relationship between the two users. Further we also determine the trustability of the content based on the content creator's acceptance measure by the social network. We use these two features in addition to the term-frequency - inverse-term-frequency match to rank the search results in context of the social network of the query-creator and provide a more specific and socially relevant result to the user.

Since the 1970's, database systems have been "compute-centric". When a computation needs the data, it requests the data, and the data are pulled through the system. We believe that this is problematic for two reasons. First, requests for data naturally incur high latency as the data are pulled through the memory hierarchy, and second, it makes it difficult or impossible for multiple queries or operations that are interested in the same data to amortize the bandwidth and latency costs associated with their data access.