No abstract

In this paper, we present a two-stage neural quality estimation model that uses multilevel task learning for translation quality estimation (QE) at the sentence, word, and phrase levels. Our approach is based on an end-to-end stacked neural model named Predictor-Estimator, which has two stages consisting of a neural word prediction model and neural QE model. To efficiently train the two-stage model, a stack propagation method is applied, thereby enabling us to jointly learn the word prediction model and QE model in a single learning mode. In addition, we deploy multilevel task learning with stack propagation, where the training examples available for all QE subtasks (i.e., sentence/word/phrase levels) are used to train a Predictor-Estimator for a specific subtask. All of our submissions to the QE task of WMT17 are ensembles that combine a set of neural models trained under different settings of varying dimensionalities and shuffling training examples, eventually achieving the best performances for all subtasks at the sentence, word, and phrase levels.

Many heterogeneous catalytic reactions occur at high temperatures, which may cause large energy costs, poor safety, and thermal degradation of catalysts. Here, we propose a light-assisted surface reaction, which catalyze the surface reaction using both light and heat as an energy source. Conventional metal catalysts such as ruthenium, rhodium, platinum, nickel, and copper were tested for CO2 hydrogenation, and ruthenium showed the most distinct change upon light irradiation. CO2 was strongly adsorbed onto ruthenium surface, forming hybrid orbitals. The band gap energy was reduced significantly upon hybridization, enhancing CO2 dissociation. The light-assisted CO2 hydrogenation used only 37% of the total energy with which the CO2 hydrogenation occurred using only thermal energy. The CO2 conversion could be turned on and off completely with a response time of only 3 min, whereas conventional thermal reaction required hours. These unique features can be potentially used for on-demand fuel production with minimal energy input.

No abstract

No abstract

As climate change becomes increasingly evident, reducing greenhouse gases including CO 2 has received growing attention. Because CO 2 is thermodynamically very stable, its conversion into value-added chemicals such as CO, CH 4 , or C 2 H 4 is difficult, and developing efficient catalysts for CO 2 conversion is important work. CO 2 can be converted using the gas-phase reaction, liquid-phase reaction, photocatalytic reaction, or electrochemical reaction. The gas-phase reaction includes the dry reforming of methane using CO 2 and CH 4 , or CO 2 hydrogenation using CO 2 and H 2. The liquid-phase reaction includes formic acid formation from pressurized CO 2 and H 2 in aqueous solution. The photocatalytic reaction is commonly known as artificial photo-synthesis, and produces chemicals from CO 2 and H 2 O under light irradiation. The electrochemical reaction can produce chemicals from CO 2 and H 2 O using electricity. In this review, the heterogeneous catalysts used for the gas-phase reaction or electrochemical reactions are discussed, because the liquid-phase reaction and photocatalytic reaction typically suffer from low productivity and poor durability. Because the gas-phase reaction requires a high reaction temperature of > 600°C, obtaining good durability is important. The strategies for designing catalysts with good activity and durability will be introduced. Various materials have been tested for electrochemical conversion, and it has been shown that specific metals can produce specific products, such as Au or Ag for CO, Sn or Bi for formate, Cu for C 2 H 4. Other unconventional catalysts for electrochemical CO 2 reduction are also introduced.

No abstract

No abstract