Most of the currently existing vision and language pre-training (VLP) methods have mainly focused on how to extract align text features. In contrast mainstream VLP methods, we highlight that two routinely applied steps during crucial impact performance pre-trained model: in-batch hard negative sampling for image-text matching (ITM) assigning large masking probability masked modeling (MLM). Afte...

This paper discusses the issue of language sampling in the development of a framework for multilingual lexical representation (MLR). The approach to MLR adopted here aims at developing a hierarchical lexical structure which permits sharing of information across languages. The multilingual nature of such a lexical hierarchy raises methodological and theoretical issues for its design and developm...

Monte Carlo model checking introduced by Smolka and Grosu is an approach to analyse non-probabilistic models using sampling and draw conclusions with a given confidence interval by applying statistical inference. Though not exhaustive, the method enables verification of complex models, even in cases where the underlying problem is undecidable. In this paper we develop Monte Carlo model checking...

We present techniques for importance sampling from distributions defined by Relational Bayesian Networks. The methods operate directly on the abstract representation language, and therefore can be applied in situations where sampling from a standard Bayesian Network representation is infeasible. We describe experimental results from using standard, adaptive and backward sampling strategies. Fur...

This paper investigates various approaches to data sampling and dimensionality reduction for discriminative language models (DLM). Being a feature based language modeling approach, the aim of DLM is to rerank the ASR output with discriminatively trained feature parameters. Using a Turkish morphology based feature set, we examine the use of online Principal Component Analysis (PCA) as a dimensio...

The “Ocean SAmpling MObile Network” (SAMON) Project is a simulation testbed for Web-based interaction among oceanographers and simulation based design of Ocean Sampling missions. In this paper, the current implementation of SAMON is presented, along with a formal model based on process algebra. Flexible optimization handles planning, mobility, evolution, and learning. A generic behavior message...

This paper introduces a new approach that directly uses latent words language models (LWLMs) in automatic speech recognition (ASR). LWLMs are effective against data sparseness because of their soft-decision clustering structure and Bayesian modeling so it can be expected that LWLMs perform robustly in multiple ASR tasks. Unfortunately, implementing a LWLM to ASR is difficult because of its comp...