Database and Related Research at IIT Kharagpur

1
Database and Related Research at IIT Kharagpur

• presented by
• P P Chakrabarti Subrato Das

2
Areas of Research

• Earlier Work
• Fuzzy Relational Databases
• ODS Object Oriented Database System
• Logic and Databases
• Search Algorithms
• Current Work
• Active Database Management Systems
• Image and Multimedia Databases
• Data Mining for Temporal Properties
• Textual Databases

3
ODS

• Supports Primitive and Complex Classes Instance
  Class Constructors Specialization (ISA
  Hierarchy) Aggregation Association (Array,
  Sets) Message and Method Rules
• Language NODAL combines procedural and
  declarative paradigm ensures encapsulation
  allows recursive rules sequential and parallel
  construct iteration

4
ODS Features

• Uses 3 valued logic for handling incomplete
  information
• Fixed Point Semantics Established
• Storage Organization Optimized
• Used in Creating a VLSI Database

5
Logic and Databases

• LogiC A language which allows calls to C
  functions in such a manner that we can have
  backtracking across the C functions
• Logi_DB A logic language which allows calls to
  both logic as well as database predicates with
  backtracking
• Constrained Systems Other than predicates and
  functions, constraints can also be specified and
  automatic constraint satisfaction can be
  performed like in CLP

6
Agent Search in limited memory

• An important problem in search in AI - how to
  search for a solution in a large implicit state
  space.
• Optimal and near optimal algorithms have been
  developed for trees with and without heuristics.
• Possible applications in large databases to be
  performed by a limited memory agent.
• Solving optimization problems in large databases
  using search methods.

7
Active Databases

• Not only support passive DBMS functions but also
  carry out actions in response to events.
• Event-Condition-Action (ECA) rules are used to
  provide modularity and timely response.
• Object Oriented ADBMS.
• Transaction Modelling and Consurrency Control in
  ADBMS.

8
Image and Multimedia Databases

• Model for Image Description
• For individual objects attributes and properties
  (O-ODDL).
• Between Objects Topological and spatial
  relationships
• Similarity Measure for Images.
• Query based on object attributes, spatial and
  topological relationships and similarity.
• Relational Engine at Backend.
• Extended to model video using temporal operators.

9
Mining for Temporal Property Identification

• Identification of state and state change.
• System Abstraction.
• Temporal Property Checking from Data with partial
  information of inherent transition system.
• Bounded as well as unbounded temporal property
  checking.

10
Other Related Areas

• Constraint Satisfaction using Multi-dimensional
  relational tables.
• Knowledge Based Systems
• Solving Constrained Optimization Problems using
  Search and Deduction.
• Knowledge based Design of Embedded Systems using
  the concept of knowledge intensive methods
• Textual Databases information retrieval,
  summarization, etc