arXiv subjects

Jing Gu

Publications and source records attributed to Jing Gu.

41 records · Page 3Linked to original sources

arXiv2020

Data Annealing for Informal Language Understanding Tasks

There is a huge performance gap between formal and informal language understanding tasks. The recent pre-trained models that improved the performance of formal language understanding tasks did not achieve a comparable result on informal language. We pro-pose a data annealing transfer learning procedure to bridge the performance gap on informal natural language understanding tasks. It successfully utilizes a pre-trained model such as BERT in informal language. In our data annealing procedure, the training set contains mainly formal text data at first; then, the proportion of the informal text data is gradually increased during the training process. Our data annealing procedure is model-independent and can be applied to various tasks. We validate its effectiveness in exhaustive experiments. When BERT is implemented with our learning procedure, it outperforms all the state-of-the-art models on the three common informal language tasks.

cs.CL↗

arXiv2020

A Tailored Pre-Training Model for Task-Oriented Dialog Generation

The recent success of large pre-trained language models such as BERT and GPT-2 has suggested the effectiveness of incorporating language priors in downstream dialog generation tasks. However, the performance of pre-trained models on the dialog task is not as optimal as expected. In this paper, we propose a Pre-trained Role Alternating Language model (PRAL), designed specifically for task-oriented conversational systems. We adopted (Wu et al., 2019) that models two speakers separately. We also design several techniques, such as start position randomization, knowledge distillation, and history discount to improve pre-training performance. We introduce a task-oriented dialog pretraining dataset by cleaning 13 existing data sets. We test PRAL on three different downstream tasks. The results show that PRAL performs better or on par with state-of-the-art methods.

cs.CL↗

arXiv2019

Quantifying Inactive Lithium in Lithium Metal Batteries

Inactive lithium (Li) formation is the immediate cause of capacity loss and catastrophic failure of Li metal batteries. However, the chemical component and the atomic level structure of inactive Li have rarely been studied due to the lack of effective diagnosis tools to accurately differentiate and quantify Li+ in solid electrolyte interphase (SEI) components and the electrically isolated unreacted metallic Li0, which together comprise the inactive Li. Here, by introducing a new analytical method, Titration Gas Chromatography (TGC), we can accurately quantify the contribution from metallic Li0 to the total amount of inactive Li. We uncover that the Li0, rather than the electrochemically formed SEI, dominates the inactive Li and capacity loss. Using cryogenic electron microscopies to further study the microstructure and nanostructure of inactive Li, we find that the Li0 is surrounded by insulating SEI, losing the electronic conductive pathway to the bulk electrode. Coupling the measurements of the Li0 global content to observations of its local atomic structure, we reveal the formation mechanism of inactive Li in different types of electrolytes, and identify the true underlying cause of low Coulombic efficiency in Li metal deposition and stripping. We ultimately propose strategies to enable the highly efficient Li deposition and stripping to enable Li metal anode for next generation high energy batteries.

cond-mat.mtrl-sci↗

arXiv2016

Self-optimizing layered hydrogen evolution catalyst with high basal-plane activity

Hydrogen is a promising energy carrier and key agent for many industrial chemical processes1. One method for generating hydrogen sustainably is via the hydrogen evolution reaction (HER), in which electrochemical reduction of protons is mediated by an appropriate catalyst-traditionally, an expensive platinum-group metal. Scalable production requires catalyst alternatives that can lower materials or processing costs while retaining the highest possible activity. Strategies have included dilute alloying of Pt2 or employing less expensive transition metal alloys, compounds or heterostructures (e.g., NiMo, metal phosphides, pyrite sulfides, encapsulated metal nanoparticles)3-5. Recently, low-cost, layered transition-metal dichalcogenides (MX2)6 based on molybdenum and tungsten have attracted substantial interest as alternative HER catalysts7-11. These materials have high intrinsic per-site HER activity; however, a significant challenge is the limited density of active sites, which are concentrated at the layer edges.8,10,11. Here we use theory to unravel electronic factors underlying catalytic activity on MX2 surfaces, and leverage the understanding to report group-5 MX2 (H-TaS2 and H-NbS2) electrocatalysts whose performance instead derives from highly active basal-plane sites. Beyond excellent catalytic activity, they are found to exhibit an unusual ability to optimize their morphology for enhanced charge transfer and accessibility of active sites as the HER proceeds. This leads to long cycle life and practical advantages for scalable processing. The resulting performance is comparable to Pt and exceeds all reported MX2 candidates.

cond-mat.mtrl-sci↗

arXiv2014

Theoretic Insight into CO2 Reduction at Active Sites of Molybdenum and Tungsten Enzymes: a π Interaction between CO2 and Tungsten Bis-Dithiolene Complexes

Active sites of molybdenum and tungsten enzymes, particularly mononuclear tungsten formate dehydrogenase (FDH) have been theoretically investigated towards their interaction with CO2. Obvious π interaction has been found between the 2e reduced metallodithiole moiety and the molecular CO2. This weak π bonding is predicated both at gas phase, noted as -6.0 kcal/mol and aqueous solvation level, -3.6 kcal/mol. Such interaction is not only limited to CO2, but also to the CO2 reduced product, i.e. formate, in the form of anion- π interaction, noted as -6.8 kcal/mol and -4.1 kcal/mol respectively in gas and aqueous solvation model. The Bailar twisted angles from 60o to 0o, governing structure preference of tungsten dithiolene from octahedron to triangle prism in their restricted structures, has been explored to evaluate such π in-terrelations with CO2 and formate. An octahedral structure with 3 kcal/mol energy lower is preferred over the triangle prismatic when such interactions are concerned.

physics.chem-ph↗