Learning the dynamics of embedded clauses

Recent work by Siegelmann has shown that the computational power of recurrent neural networks matches that of Turing Machines. One important implication is that complex language classes (infinite languages with embedded clauses) can be represented in neural networks. Proofs are based on a fractal encoding of states to simulate the memory and operations of stacks. In the present work, it is shown that similar stack-like dynamics can be learned in recurrent neural networks from simple sequence prediction tasks. Two main types of network solutions are found and described qualitatively as dynamical systems: damped oscillation and entangled spiraling around fixed points. The potential and limitations of each solution type are established in terms of generalization on two different context-free languages. Both solution types constitute novel stack implementations - generally in line with Siegelmann's theoretical work - which supply insights into how embedded structures of languages can be handled in analog hardware.

Tool support for testing concurrent Java components

Concurrent programs are hard to test due to the inherent nondeterminism. This paper presents a method and tool support for testing concurrent Java components. Too[ support is offered through ConAn (Concurrency Analyser), a too] for generating drivers for unit testing Java classes that are used in a multithreaded context. To obtain adequate controllability over the interactions between Java threads, the generated driver contains threads that are synchronized by a clock. The driver automatically executes the calls in the test sequence in the prescribed order and compares the outputs against the expected outputs specified in the test sequence. The method and tool are illustrated in detail on an asymmetric producer-consumer monitor. Their application to testing over 20 concurrent components, a number of which are sourced from industry and were found to contain faults, is presented and discussed.

An approach to verifying and debugging simulation models governed by ordinary differential equations: Part 1. Methodology for residual generation

For dynamic simulations to be credible, verification of the computer code must be an integral part of the modelling process. This two-part paper describes a novel approach to verification through program testing and debugging. In Part 1, a methodology is presented for detecting and isolating coding errors using back-to-back testing. Residuals are generated by comparing the output of two independent implementations, in response to identical inputs. The key feature of the methodology is that a specially modified observer is created using one of the implementations, so as to impose an error-dependent structure on these residuals. Each error can be associated with a fixed and known subspace, permitting errors to be isolated to specific equations in the code. It is shown that the geometric properties extend to multiple errors in either one of the two implementations. Copyright (C) 2003 John Wiley Sons, Ltd.

Enhanced effector and memory CTL responses generated by incorporation of receptor activator of NF-kappa B(RANK)/RANK ligand costimulatory molecules into dendritic cell immunogens expressing a human tumor-specific antigen

The outcome of dendritic cell (DC) presentation of Ag to T cells via the TCR/MHC synapse is determined by second signaling through CD80/86 and, importantly, by ligation of costimulatory ligands and receptors located at the DC and T cell surfaces. Downstream signaling triggered by costimulatory molecule ligation results in reciprocal DC and T cell activation and survival, which predisposes to enhanced T cell-mediated immune responses. In this study, we used adenoviral vectors to express a model tumor Ag (the E7 oncoprotein of human papillomavirus 16) with or without coexpression of receptor activator of NF-kappaB (RANK)/RANK ligand (RANKL) or CD40/CD40L costimulatory molecules, and used these transgenic DCs to immunize mice for the generation of E7-directed CD8(+) T cell responses. We show that coexpression of RANK/RANKL, but not CD40/CD40L, in E7-expressing DCs augmented E7-specific IFN-gamma-secreting effector and memory T cells and E7-specific CTLs. These responses were also augmented by coexpression of T cell costimulatory molecules (RANKL and CD40L) or DC costimulatory molecules (RANK and CD40) in the E7-expressing DC immunogens. Augmentation of CTL responses correlated with up-regulation of CD80 and CD86 expression in DCs transduced with costimulatory molecules, suggesting a mechanism for enhanced T cell activation/survival. These results have generic implications for improved tumor Ag-expressing DC vaccines, and specific implications for a DC-based vaccine approach for human papillomavirus 16-associated cervical carcinoma.