We present a new open source toolkit for phrase-based and syntax-based machine translation. The toolkit supports several state-of-the-art models developed in statistical machine translation, including the phrase-based model, the hierachical phrase-based model, and various syntaxbased models. The key innovation provided by the toolkit is that the decoder can work with various grammars and offers...

AVS1-P2 is the newest video standard of Audio Video coding Standard (AVS) workgroup of China, which provides close performance to H.264/AVC main profile with lower complexity. In this paper, a platform-independent software package with macroblockbased (MB-based) architecture is proposed to facilitate AVS video standard implementation on embedded system. Compared with the frame-based architectur...

Context Adaptive Variable Length Decoding (CAVLD) module takes the lion chair of the H.264/AVC video decoder time due to its complexity. In order to ameliorate decoding speed, a new CAVLD algorithm and an efficient internal memory design were implemented on Digital Signal Processor (DSP). The proposed CAVLD algorithm, Zero Length Prefix (ZLP), was designed to optimize the first syntax element: ...

JPEG-2000 is an emerging standard for still image compression. This paper provides a brief history of the JPEG-2000 standardization process, an overview of the standard, and some description of the capabilities provided by the standard. Part I of the JPEG-2000 standard specifies the minimum compliant decoder, while Part II describes optional, value-added extensions. Although the standard specif...

Neural machine translation (NMT) models are able to partially learn syntactic information from sequential lexical information. Still, some complex syntactic phenomena such as prepositional phrase attachment are poorly modeled. This work aims to answer two questions: 1) Does explicitly modeling target language syntax help NMT? 2) Is tight integration of words and syntax better than multitask tra...

Neural machine translation (NMT) models are able to partially learn syntactic information from sequential lexical information. Still, some complex syntactic phenomena such as prepositional phrase attachment are poorly modeled. This work aims to answer two questions: 1) Does explicitly modeling target language syntax help NMT? 2) Is tight integration of words and syntax better than multitask tra...

In this paper we present superficially a software tool ---a DVB-MHP compliant Transport Stream Generator--designed for being part of the testbed of a MHP decoder prototype. The tool was implemented partially making use of XML technology and the software reuse philosophy (free-distributed parsers and tools) that in a sense goes with this special syntax.

The H.264/AVC is the most recent standard of video compression/decompression for future broadband network. This standard was developed through the Joint Video Team (JVT) from the ITU-T Video Coding Experts Group and the ISO/IEC MPEG standardization committee. In this project H.264 decoder functional block such as Context based Binary arithmetic coding (CABAC), Inverse Quantization and Inverse D...

Tasks like code generation and semantic parsing require mapping unstructured (or partially structured) inputs to well-formed, executable outputs. We introduce abstract syntax networks, a modeling framework for these problems. The outputs are represented as abstract syntax trees (ASTs) and constructed by a decoder with a dynamically-determined modular structure paralleling the structure of the o...

End-to-end training makes the neural machine translation (NMT) architecture simpler, yet elegant compared to traditional statistical machine translation (SMT). However, little is known about linguistic patterns of morphology, syntax and semantics learned during the training of NMT systems, and more importantly, which parts of the architecture are responsible for learning each of these phenomena...