Tuple

About: Tuple is a research topic. Over the lifetime, 6513 publications have been published within this topic receiving 146057 citations. The topic is also known as: tuple & ordered tuplet.

In Search of the Lost Schema

Abstract: We study the problem of rediscovering the schema of nested relations that have been encoded as strings for storage purposes. We consider various classes of encoding functions, and consider the markup encodings, which allow to find the schema without knowledge of the encoding function, under reasonable assumptions on the input data. Depending upon the encoding of empty sets, we propose two polynomial on-line algorithms (with different buffer size) solving the schema finding problem. We also prove that with a high probability, both algorithms find the schema after examining a fixed number of tuples, thus leading in practice to a linear time behavior with respect to the database size for wrapping the data. Finally, we show that the proposed techniques are well-suited for practical applications, such as structuring and wrapping HTML pages and Web sites.

Bipolar Queries Using Various Interpretations of Logical Connectives

Abstract: In [1,2] we studied various concepts of bipolar queries (cf. Dubois and Prade [3]). We advocated the use of a fuzzified version of the original crisp definition by Lacroix and Lavency [4]. However, the general fuzzification proposed leaves open the choice of a representation of logical connectives and quantifiers. In the present paper we study the influence of the choice some representations that are popular in fuzzy logic on matching degrees of the tuples and their resulting ordering.

Learned Cardinalities: Estimating Correlated Joins with Deep Learning

Abstract: We describe a new deep learning approach to cardinality estimation. MSCN is a multi-set convolutional network, tailored to representing relational query plans, that employs set semantics to capture query features and true cardinalities. MSCN builds on sampling-based estimation, addressing its weaknesses when no sampled tuples qualify a predicate, and in capturing join-crossing correlations. Our evaluation of MSCN using a real-world dataset shows that deep learning signiicantly enhances the quality of cardinality estimation, which is the core problem in query optimization.

Sampling over joins for database systems

Abstract: A database server supports weighted and unweighted sampling of records or tuples in accordance with desired sampling semantics such as with replacement (WR), without replacement (WoR), or independent coin flips (CF) semantics, for example. The database server may perform such sampling sequentially not only to sample non-materialized records such as those produced as a stream by a pipeline in a query tree for example, but also to sample records, whether materialized or not, in a single pass. The database server also supports sampling over a join of two relations of records or tuples without requiring the computation of the full join and without requiring the materialization of both relations and/or indexes on the join attribute values of both relations.

Coordination and Access Control in Open Distributed Agent Systems: The TuCSoN Approach

Abstract: Coordination and access control are related issues in open distributed agent systems, being both concerned with governing interaction between agents and resources. In particular, while coordination deals with enabling interaction and making it fruitful, access control is meant to control interaction to make it harmless. We argue that this twofold facet has to be supported by a system in a uniform and decentralised manner. To this end, we describe how the application of the TuCSoN tuple-based coordination model over a hierarchical topology is well-suited in this context. On the one hand, policies can be enforced by means of a single mechanism based on tuples and can be scoped to manage access to groups of distributed resources. On the other hand, agents can interact along a hierarchical infrastructure by applying a standard tuple-based communication template. This makes TuCSoN a single coherent framework for the design and development of Internet-based multiagent systems, which takes coordination as the basis for dealing with network topology and access control in a uniform way.