We introduce a general strategy for improving neural sequence generation by incorporating knowledge about the future. Our decoder combines a standard sequence decoder with a ‘soothsayer’ prediction function Q that estimates the outcome in the future of generating a word in the present. Our model draws on the same intuitions as reinforcement learning, but is both simpler and higher performing, a...

We propose a conditional non-autoregressive neural sequence model based on iterative refinement. The proposed model is designed based on the principles of latent variable models and denoising autoencoders, and is generally applicable to any sequence generation task. We extensively evaluate the proposed model on machine translation (En↔De and En↔Ro) and image caption generation, and observe that...

Next-generation sequencing technologies enable the rapid cost-effective production of sequence data. To evaluate the performance of these sequencing technologies, investigation of the quality of sequence reads obtained from these methods is important. In this study, we analyzed the quality of sequence reads and SNP detection performance using three commercially available next-generation sequenc...

This paper describes a method for automatically generating Unique Input Output (UIO) sequences for FSM conformance testing. UIOs are used in conformance testing to verify the end state of a transition sequence. UIO sequence generation is represented as a search problem and genetic algorithms are used to search this space. The empirical evidence indicates that the proposed method yields consider...

Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of sequence-to-sequence deep learning technologies such as Transformer-based language models. This advancement led more fluent and coherent NLG, leading downstream tasks abstractive summarization, dialogue generation data-to-text generation. However, it is also apparent that based prone halluc...

A new fractional accumulation technique based on discrete sequence convolution transform was developed. The system, whose unit impulse response is the sequence, constructed; then, order extended to orders. accumulative grey forecasting model GMr*(1,1) established convolution. From viewpoint of convolution, we can better understand mechanism generation. Real cases were used verify validity and e...

Logical Natural Language Generation, i.e., generating textual descriptions that can be logically entailed by a table, has been challenge due to the low fidelity of generation. Previous works have addressed this problem annotating interim logical programs control generation contents and semantics, presented task table-aware form text (Logic2text) However, although table instances are abundant in...

Proteins are sources of many peptides with diverse biological activity. Some of them are considered as valuable components of foods and drug targets with desired and designed biological activity. We are now entering an era rich in biological data in which the field of bioinformatics is poised to exploit this information in increasingly powerful ways. There are currently many databases all over ...

-A new key generation technique for high speed communication is presented. Starting with a short random sequence it generates a pseudo random sequence over {0,1} with good properties for a one time pad encryption. Since this generation is based on extremely simple boolean operations it allows for even higher data rates than standard symmetric crypto algorithms.