This paper describes a novel method for computing a consensus translation from the outputs of multiple machine translation (MT) systems. The outputs are combined and a possibly new translation hypothesis can be generated. Similarly to the well-established ROVER approach of (Fiscus, 1997) for combining speech recognition hypotheses, the consensus translation is computed by voting on a confusion ...

This paper describes a novel method for computing a consensus translation from the outputs of multiple machine translation (MT) systems. The outputs are combined and a possibly new translation hypothesis can be generated. Similarly to the well-established ROVER approach of (Fiscus, 1997) for combining speech recognition hypotheses, the consensus translation is computed by voting on a confusion ...

In general, there are two ways to deal with one-against-all multiclass neural network classification. The first way is the use of a single k-class neural network trained with multiple outputs. Another way is the use of multiple binary neural networks. This paper focuses on the later way in which multiple complementary neural networks are applied to one-against-all instead of using only multiple...

Previous work has shown that clock skew can be considered an identifying feature of a physical system, independent of time of measurement, network location of subject and observer, and other factors. Under certain reasonable assumptions, it is possible to fingerprint, or uniquely identify within some anonymity set, a remote machine by observing some of its network traffic and observing changes ...

Tensor network algorithms seek to minimize correlations compress the classical data representing quantum states. and similar tools—called tensor methods—form backbone of modern numerical methods used simulate many-body physics have a further range applications in machine learning. Finding contracting states is computational task, which may be accelerated by computing. We present algorithm that ...

The input-queued switching architecture is becoming an attractive alternative for high speed switches owing to its scalability. Tremendous eeorts have been made to overcome the throughput problem caused by the contentions occurred at input and output sides of a switch. Existing input queueing algorithms mostly aim at improving through-put without considering QoS features. In this paper, a new a...

Bipartite network flow problems naturally arise in applications such as selective assembly and preemptive scheduling. This paper presents fast algorithms for these problems that take advantage of special properties of the associated bipartite networks. We show a connection between selective assembly and the Earliest Due Date (EDD) scheduling rule, and we show that EDD can be implemented in line...

Modern studies in the field of environment science and engineering show that deterministic models struggle to capture the relationship between the concentration of atmospheric pollutants and their emission sources. The recent advances in statistical modeling based on machine learning approaches have emerged as solution to tackle these issues. It is a fact that, input variable type largely affec...

We propose new heuristics along with an augmented-neural-network (AugNN) formulation for solving the makespan minimization task-scheduling problem for the non-identical machine environment. We explore four task and three machine-priority rules, resulting in 12 combinations of single-pass heuristics. The task-priority rules are Highest-LevelFirst (HLF), Highest-Total-Remaining-Processing-Time-Fi...

Recently, machine learning techniques have been actively applied to the identification of Alzheimer’s disease (AD) and mild cognitive impairment (MCI). However, most of the existing methods focus on using only single network property, although combination of multiple network properties such as local connectivity and topological properties may be more powerful. Employing the kernel-based method,...