UNM Computer Science

Colloquia Archive - Fall 2003

64-bit PowerPC port of Jikes Research Virtual Machine and memory management in 64-bit space

Date: Thursday, December 11th
Time: 11am-12:15pm
Location: Woodward 149

Sergiy Kyrylkov, <sergiy@cs.unm.edu>
Department of Computer Science, UNM

Abstract:
64-bit computing, contrasted with 32-bit computing, can be characterized by very large memory support, very large application virtual address spaces, and 64-bit integer computation, using 64-bit general purpose registers. In such 64-bit systems, an application's virtual address space is measured in terabytes. At the same time, there is an increasing number of programs that can exploit this opportunity. Some applications can exhibit performance gains only when the physical memory available is large enough. On the other hand, other applications, including virtual machines with new memory management algorithms, can potentially benefit merely from having a very large virtual address space without additional physical memory. I will describe a 64-bit PowerPC port of Jikes Research Virtual Machine and design and implementation of a 64-bit Older-First collector and allocator with address-order write barriers.

Human-Like Cognitive Models as a Basis for Cognitive Systems

Date: Tuesday, December 9th
Time: 11am-12:15pm
Location: Woodward 149

J Chris Forsythe, <jcforsy@sandia.gov>
Sandia National Laboratories

Abstract:
This presentation will summarize an ongoing program of research and development at Sandia National Laboratories focused on development of cognitive systems technologies. The approach being taken is unique in its emphasis on a neurologically-inspired, psychologically plausible computational model of human cognition. This model will be briefly discussed, as well as the process by which the model has been designed so as to justify claims for psychological plausibility. The presentation will also address cognitive system concepts that utilize this modeling framework. One example employs a real-time model of an operator to enable detection of potential errors and facilitate error recovery. A second example utilizes expert cognitive mdoels for surveillance of data sources and includes features in which individual models may collaborate with one another. Finally, there will be discussion of current efforts concerned with automated knowledge capture and development of an individualized digital aide.

Spatio-temporal Covariance Model for Human Brain MEG Analysis

Date: Thursday, December 4th
Time: 11am-12:15pm
Location: Woodward 149

Sergey Plis, <pliz@cs.unm.edu>
Department of Computer Science, UNM

Abstract:
There are many problems in data analysis or signal processing in which one wants to discern what is different from a "signal" state and a "background" state. This is also true in the brain mapping modality of Magnetoencephalography (MEG) in which one measures the magnetic field outside a person's head generated from the electrical currents from active sets of brain neurons. Here the signal state is the brain's response to a given stimulus and the background state is ongoing neural activity not time-locked to that stimulus. The ability to characterize the statistical nature of the background activity is important for correctly inferring the brain's response. This background is sufficiently complex and correlated that it is not practical to construct its full covariance using conventional means. We present a model for the spatiotemporal noise covariance that has a small enough number of free parameters to be estimated correctly from reasonable amounts of data. In addition this model has the property that the inverse of the covariance is easy to calculate, which is important for using this covariance in an analysis. Our approach models data with spatiotemporal dependencies better than an other approach (also to be presented) and is only linearly more complex in the number of parameters.

Automated Reasoning and Symbolic Computation

Date: Thursday, November 20th
Time: 11am-12:15pm
Location: Woodward 149

Deepak Kapur, <kapur@cs.unm.edu>
Department of Computer Science, UNM

Abstract:
I will discuss my current research in automated reasoning and symbolic computation. I will talk about two topics: (i) the use
of decision procedures for quantifier-free theories to help mechanize induction, and (ii) Dixon resultants for solving polynomial equations. Time permitting, new research on a method for computing invariants of loop programs inspired by Colon, Sankaranarayanan and Sipma (CAV 2003) will be presented.