Stan C. Kwasny

Learn More
As a programming language for computational linguistics, Prolog is a relative newcomer. Mark Wallace, however, demonstrates very clearly in this timely book the value of this important tool, especially as it relates to the building of natural language front ends and interfaces to database systems. He takes a very practical approach which should appeal to(More)
Representation poses important challenges to connectionism. The ability to structurally compose representations is critical in achieving the capability considered necessary for cognition. We are investigating distributed patterns that represent structure as part of a larger effort to develop a natural language processor. Recursive Auto-Associative Memory(More)
RATIONALE AND OBJECTIVES Use of a neural network to diagnose focal lesions of bone was evaluated. METHODS Imaging features of 709 lesions were encoded into a predetermined database. Data were divided into four groups and were analyzed using cross-validation by a two-layer feed-forward neural network. RESULTS The lesions comprised 43 different pathologic(More)
RATIONALE AND OBJECTIVES The authors evaluated the feasibility of combining wavelet transform and artificial neural network (ANN) technologies to prescreen mammograms for masses. METHODS AND MATERIALS Fifty-five mammograms (29 with masses and 26 without) were digitized to 100-mm resolution and processed by using wavelet transformation. These wavelets were(More)
RATIONALE AND OBJECTIVES We evaluated the potential for a neural network to screen candidates for emergency cranial computed tomography (CT) scans in an emergency department setting. METHODS Data were collected from 1625 patients undergoing emergency cranial CT scanning in two different emergency departments (EDs). Singular value decomposition (SVD) was(More)
People can differentiate spoken languages without understanding them, and, in some sense, this differentiation can only be done without understanding the language. When we consider a multi-lingual person trying to understand an utterance spoken in one of the languages with which they are familiar, they will first decide which language this utterance belongs(More)
TRAINREC is a system for training feedforward and recurrent neural networks that incorporates several ideas. It uses the conjugate-gradient method which is demonstrably more efficient than traditional backward error propagation. We assume epoch-based training and derive a new error function having several desirable properties absent from the traditional(More)