Abstract: Some links are omitted from indexes associated with semantic network elements and clustering is influenced by such omissions.

Abstract: An inference system responsive to omitting links from indexes in semantic network elements is described. The inference system prioritizes strategies that avoid trying to follow links from an index from which they may have been omitted.

Abstract: A method, computing system, and computer program product are provided for quickly and space-efficiently mapping an object's address to its home node in a computing system with a very large (possibly multi-petabyte) data set. The addresses of objects comprise three fields: a chunk number, a region sub-index within the chunk, and an offset within the region, with chunks being used to achieve good compromise between small lookup tables and reducing waste of usable virtual address space.

Abstract: A garbage collector is disclosed that permits extensive separation of mutators and the garbage collector from a synchronization perspective. This relative decoupling of mutator and collector operation allows the garbage collector to perform relatively time-intensive operations during garbage collection without substantially slowing down mutators. The present invention makes use of this flexibility by first conservatively determining which objects in a set of regions of interest are live, then planning where to copy the objects (preferably including clustering), and finally performing the actual copying.

Abstract: In an object-relocating garbage collector, objects are copied and new copies taken into use concurrently with mutator execution without needing to use a read barrier, and importantly, without requiring verification (read-back) of each copied word and without requiring atomic instructions for the copying. Write barriers, thread-local write barrier buffers and processing them by the garbage collector, and write propagation are used for achieving this.

Abstract: A fast method for determining which actions to take in a write barrier in a concurrent garbage collector is described. A region status array indexed by a region index computed from the written address is used for determining the status of the region containing the written object and for selecting, in part, the actions taken by the write barrier. By carefully manipulating the region status array, various operations and changes in write barrier actions can be performed very efficiently.

Abstract: Objects are copied concurrently with mutator execution, while tracking writes to the objects being copied. Objects (or fields) that are written into during copying are re-copied to the same destination locations. Mutators use the original objects until copying is complete and are, in some embodiments, atomically (with respect to the mutators) switched to use the new copies, together with a final re-copy.

Abstract: A write barrier is implemented using thread-local hash table based write barrier buffers. The write barrier, executed by mutator threads, stores addresses of written memory locations or objects in the thread-local hash tables, and during garbage collection, an explicit or implicit union of the addresses in each hash table is used in a manner that is tolerant to an address appearing in more than one hash table.

Abstract: During garbage collection, writes to objects being copied (relocated) are monitored (tracked) using a write barrier that uses a thread-local write barrier buffer. In the preferred embodiment, soft synchronization is used for reading the thread-local write barrier buffers. In response to detecting a write to an object, the object may be re-copied, the copying may be made to fail, the write may be propagated to another copy of the object, or, e.g., another node in a distributed system may be notified of the write.

Abstract: A method and computer system for determining the memory address range of a subtree of a linearized tree of objects or the original version of such tree when the tree has been modified or is concurrently being modified by other threads. The primary application of the method is in garbage collection for multiprocessor shared memory computers with very large memories, but it is also relevant for a number of other applications.

Abstract: In many reference resolution problems there are many candidate referents, and the overhead of enumerating them can be considerable. The overhead is reduced by stopping enumeration before all candidate referents have been enumerated, utilizing the properties of ordered and semi-ordered enumerators. Converting semi-ordered enumerators into ordered enumerators and combining several ordered enumerators into a single using dynamic weightings for handling determiner interpretations are disclosed.

Abstract: A method of encoding data structures using compressed object encodings during serialization. A compressed representation of the data is generated directly while encoding. Serialization means converting a data structure to a string of bytes for external storage or communication.

Abstract: Ambiguities in a natural language expression are interpreted by jointly disambiguating multiple alternative syntactic and semantic interpretations. More than one syntactic alternative, represented by parse contexts, are analyzed together with joint analysis of referents, word senses, relation types, and layout of a semantic representation for each syntactic alternative. Best combinations of interpretations are selected from all participating parse contexts, and are used to form parse contexts for the next step in parsing.

Abstract: At least two ambiguous aspects of the meaning of a natural language expression are disambiguated jointly. In the preferred embodiment, word sense ambiguity, reference ambiguity, and relation ambiguity are resolved simultaneously, finding the disambiguation result(s) that simultaneously optimize the weight of the solution, taking into account semantic information, constraints, and common sense knowledge. Choices are enumerated for each constituent being disambiguated, combinations of choices are constructed and evaluated according to semantic information on which meanings are sensible, and the choices with the best weights are selected, with the enumeration pruned aggressively to reduce computational cost.

Abstract: Disambiguation of the meaning of a natural language expression proceeds by constructing a natural language expression, and then incrementally specializing the meaning representation to more specific meanings as more information and constraints are obtained, in accordance with one or more specialization hierarchies between semantic descriptors. The method is generalized to disjunctive sets of interpretations that can be specialized hierarchically.

Abstract: Movable and elliptic constituents are handled in a parser by inserting synthetic tokens that do not occur in the input. Parser actions can push a syntax tree or semantic value to be realized later as a synthetic token, and some synthetic tokens (for cataphoric ellipsis) may be inserted without a prior push but require a later definition. At clause boundary it may be checked that all mandatory tokens have been inserted.

Abstract: Indexing overhead in a semantic network is reduced by omitting some links from indexes in semantic network elements in response to one or more omission rules. Inference strategy over the semantic network is modified to prioritize inference strategies in such a way that the impact of the omissions on inference results is minimized.

Abstract: Concurrency control overhead in transactional memory and main memory databases is reduced by automatically selecting the appropriate isolation level for each operation based on the objects accessed by the operation.

Abstract: Computer system, method and article of manufacture for garbage collection in computer systems that combines individual objects into linearized trees called multiobjects and thereafter utilizes multiobjects to implement most memory management operations such as detection of dead objects and compaction without the need to inspect individual objects, thereby improving performance and reducing power consumption.

Abstract: A consistent snapshot of a large main memory knowledge base is saved to persistent storage without blocking the application for the duration of serializing and writing the knowledge base. Taking the snapshot comprises bringing the knowledge base to a consistent state (in a multithreaded application), using virtual memory facilities to obtain a copy-on-write copy of the knowledge base in memory, and using a separate thread or process to serialize the copy-on-write copy to persistent storage.