Arrow Research search

Author name cluster

Paul Cohen

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

11 papers
2 author rows

Possible papers

11

IJCAI Conference 2011 Conference Paper

Activity Recognition with Finite State Machines

  • Wesley Kerr
  • Anh Tran
  • Paul Cohen

This paper shows how to learn general, Finite State Machine representations of activities that function as recognizers of previously unseen instances of activities. The central problem is to tell which differences between instances of activities are unimportant and may be safely ignored for the purpose of learning generalized representations of activities. We develop a novel way to find the "essential parts" of activities by a greedy kind of multiple sequence alignment, and a method to transform the resulting alignments into Finite State Machine that will accept novel instances of activities with high accuracy.

PDF Details DOI

IJCAI Conference 2009 Conference Paper

  • Daniel Hewlett
  • Paul Cohen

BOOTSTRAP VOTING EXPERTS (BVE) is an extension to the VOTING EXPERTS algorithm for unsupervised chunking of sequences. BVE generates a series of segmentations, each of which incorporates knowledge gained from the previous segmentation. We show that this method of bootstrapping improves the performance of VOTING EXPERTS in a variety of unsupervised word segmentation scenarios, and generally improves both precision and recall of the algorithm. We also show that Minimum Description Length (MDL) can be used to choose nearly optimal parameters for VOTING EX- PERTS in an unsupervised manner.

PDF Details

IJCAI Conference 2005 Conference Paper

Transfer in Learning by Doing

  • Bill Krueger
  • Tim Oates
  • Tom Armstrong
  • Paul Cohen
  • Carole

We develop two related themes, learning procedures and knowledge transfer. This paper introduces two methods for learning procedures and one for transferring previously-learned knowledge to a slightly different task. We demonstrate by experiment that transfer accelerates learning.

PDF

AAAI Conference 1999 Conference Paper

Does Prior Knowledge Facilitate the Development of Knowledge-Based Systems?

  • Paul Cohen
  • University of Massachusetts; Vinay Chaudhri
  • SRI International; Adam Pease
  • Teknowledge; Robert Schrag
  • IET Inc.

One factor that affects the rate of knowledge base construction is the availability and reuse of prior knowledge in ontologies and domain-specific knowledge bases. This paper reports an empirical study of reuse performed in the first year of the High Performance Knowledge Bases (HPKB) initiative. The study shows that some kinds of prior knowledge help more than others, and that several factors affect how much use is made of the knowledge.

PDF Details

AAAI Conference 1999 Conference Paper

What Are Contentful Mental States? Dretske’s Theory of Mental Content Viewed in the Light of Robot Learning and Planning Algorithms

  • Paul Cohen
  • University of Massachusetts
  • Mary Litch
  • University of Alabama

One concern of philosophy of mind is how sensorimotor agents such as human infants can develop contentful mental states. This paper discusses Fred Dretske’s theory of mental content in the context of results from our work with mobile robots. We argue that Dretske’s theory, while attractive in many ways, relies on a distinction between kinds of representations that cannot be practically maintained when the subject of one’s study is robotic agents. In addition, Dretske fails to distinguish classes of representations that carry different kinds of mental content. We conclude with directions for a theory of mental content that maintains the strengths of Dretske’s theory.

PDF Details

ICRA Conference 1992 Conference Paper

Automatic detection of three-dimensional solder defects using a brightness-based approach

  • L. Bérard
  • Paul Cohen
  • N. Begnoche

Addresses the problem of detecting a specific type of solder defect in IC mounts, namely excessively high solder (EHS) points. First, solder points which were suspected to be EHS points were preselected by means of a reflective method, which used the relationship between pixel intensities and surface gradients under a specially designed lighting system. By making assumptions about the local surface curvature the surface local orientation can be determined from a single brightness image. Surface reconstruction was obtained for the top region of the solder point. EHS points were then detected by using the correlation between top surface geometry and solder height. >

Details

IJCAI Conference 1991 Conference Paper

Shading-Based Two-View Matching

  • Michel Audette
  • Paul Cohen
  • Juyang Weng

This paper presents a region-based stereo matching algorithm, in which the regions are computed from the Gaussian (K) and mean (H) curvature information of the image intensity surface. A region is defined as a connected area of constant K/H sign combination, and thus constitutes the footprint of a local section of the intensity surface, whose shape is a peak, a pit, a positive or a negative saddle. Region adjacency information is explicited by means of a Voronoi graph representation of the region map. Matching between nodes of identical shape types in the two region maps is established by comparing the topological configuration of their immediate neighborhoods. The disparity map is established through a coarseto-fine strategy.

PDF