Search

Collaborative Filtering (CF) models, despite their great success, suffer from severe performance drops due to popularity distribution shifts, where these changes are ubiquitous and inevitable in real-world scenarios. Unfortunately, most leading popularity debiasing strategies, rather than tackling the vulnerability of CF models to varying popularity distributions, require prior knowledge of the test distribution to identify the degree of bias and further learn the popularity-entangled representations to mitigate the bias. Consequently, these models result in significant performance benefits in the target test set, while dramatically deviating the recommendation from users' true interests without knowing the popularity distribution in advance. In this work, we propose a novel learning framework, Invariant Collaborative Filtering (InvCF), to discover disentangled representations that faithfully reveal the latent preference and popularity semantics without making any assumption about the popularity distribution. At its core is the distillation of unbiased preference representations (i.e., user preference on item property), which are invariant to the change of popularity semantics, while filtering out the popularity feature that is unstable or outdated. Extensive experiments on five benchmark datasets and four evaluation settings (i.e., synthetic long-tail, unbiased, temporal split, and out-of-distribution evaluations) demonstrate that InvCF outperforms the state-of-the-art baselines in terms of popularity generalization ability on real recommendations. Visualization studies shed light on the advantages of InvCF for disentangled representation learning. Our codes are available at https://github.com/anzhang314/InvCF.

Two-dimensional (2D) materials have been developed into various catalysts with high performance, but employing them for developing highly stable and active nonprecious hydrogen evolution reaction (HER) catalysts still encounters many challenges. To this end, the machine learning (ML) screening of HER catalysts is accelerated by using genetic programming (GP) of symbolic transformers for various typical 2D MA2Z4 materials. The values of the Gibbs free energy of hydrogen adsorption (ΔGH*) are accurately and rapidly predicted via extreme gradient boosting regression by using only simple GP-processed elemental features, with a low predictive root-mean-square error of 0.14 eV. With the analysis of ML and density functional theory (DFT) methods, it is found that various electronic structural properties of metal atoms and the p-band center of surface atoms play a crucial role in regulating the HER performance. Based on these findings, NbSi2N4 and VSi2N4 are discovered to be active catalysts with thermodynamical and dynamical stability as ΔGH* approaches to zero (-0.041 and 0.024 eV). In addition, DFT calculations reveal that these catalysts also exhibit good deuterium evolution reaction (DER) performance. Overall, a multistep workflow is developed through ML models combined with DFT calculations for efficiently screening the potential HER and DER catalysts from 2D materials with the same crystal prototype, which is believed to have significant contribution to catalyst design and fabrication.

In this study, machine learning (ML) models combined with density functional theory (DFT) calculations and Gibbs free energy of hydrogen adsorption (ΔGH*) were employed to facilitate the high-throu...

Two-dimensional (2D) materials have been developed into various catalysts with high performance, but employing them for developing highly stable and active nonprecious hydrogen evolution reaction (HER) catalysts still encounters many challenges. To this end, the machine learning (ML) screening of HER catalysts is accelerated by using genetic programming (GP) of symbolic transformers for various typical 2D MA

Poisoning devitalization of Pt catalyst caused by the absorption of carbon monoxide is an important issue in direct Methanol Fuel Cell (DMFC). To solve this problem, this work introduced a novel nano-structured Pt catalytic electrode, in which RuO2 modified TiO2 nanotube arrays (TiO2-NTs) was used as a carrier for the load of Pt nanocatalysts. Specifically, RuCl3 sol was filled into the voids of TiO2-NTs under vacuum condition, followed by thermal decomposition to form RuO2/TiO2-NTs support, and then Pt particles were loaded on the RuO2/TiO2-NTs support by pulse potential electrodeposition from H2PtCl6 aqueous solution. The electrochemical results show that the methanol oxidation current on Pt/RuO2/TiO2-NTs is much higher than that on Pt/TiO2-NTs. In addition, the current attenuation on Pt/RuO2/TiO2-NTs with the increased scan cycle is also decreased. The Pt/RuO2/TiO2-NTs electrode with 8 g m−2 RuO2 exhibits the most stable performance, indicating a strong effects of anti CO poisoning endowed by RuO2. In Nyquist diagrams, one capacitance arc representing the action of deprivation of H atom appears in the first quadrant and one inductance arc representing the action of deprivation of CO appears in the fourth quadrant. From the fitting results, both the reaction resistance Rct and the inductance L decrease with the argument of RuO2 content under bias potential of 600 mV, and in this case CO oxidation is the rate controlling step.

Academic knowledge services have substantially facilitated the development of the science enterprise by providing a plenitude of efficient research tools. However, many applications highly depend on ad-hoc models and expensive human labeling to understand scientific contents, hindering deployments into real products. To build a unified backbone language model for different knowledge-intensive academic applications, we pre-train an academic language model OAG-BERT that integrates both the heterogeneous entity knowledge and scientific corpora in the Open Academic Graph (OAG) -- the largest public academic graph to date. In OAG-BERT, we develop strategies for pre-training text and entity data along with zero-shot inference techniques. In OAG-BERT, we develop strategies for pre-training text and entity data along with zero-shot inference techniques. Its zero-shot capability furthers the path to mitigate the need of expensive annotations. OAG-BERT has been deployed for real-world applications, such as the reviewer recommendation function for National Nature Science Foundation of China (NSFC) -- one of the largest funding agencies in China -- and paper tagging in AMiner. All codes and pre-trained models are available via the CogDL toolkit., Accepted to KDD 2022

The electrocatalytical process is the most efficient way to produce ammonia (NH

Machine learning (ML) models combined with density functional theory (DFT) calculations are employed to screen and design hydrogen evolution reaction (HER) catalysts from various bare and single-atom doped MBenes materials. The values of Gibbs free energy of hydrogen adsorption (ΔGH*) are accurately predicted via support vector algorithm only by using simply structural and elemental features. With the analysis of combined descriptors and the feature importance, the Bader charge transfer of surface metal is a key factor to influence HER activity of MBenes. Co/Ni2B2, Pt/Ni2B2, Co2B2, Os/Co2B2 and Mn/Co2B2 are screened from 271 MBenes and MXenes as active catalysts, with the near-zero ΔGH* of 0.089, −0.082, −0.13, −0.087 and −0.044 eV, respectively. Finally, stable Co2B2 and Mn/Co2B2 are considered as the excellent HER catalysts due to |ΔGH*

The electrocatalytical process is the most efficient way to produce ammonia (NH3) under ambient conditions, but developing a highly efficient and low-cost metal-free electrocatalysts remains a major scientific challenge. Hence, single atom and double boron (B) atoms doped 2D graphene-like carbon nitride (C2N-h2D) electrocatalysts have been designed (B@C2N and B2@C2N), and the efficiency of N2 reduction reaction (NRR) is examined by density functional theory calculation. The results show that the single and double B atoms can both be strongly embedded in natural nanoporous C2N with superior catalytic activity for N2 activation. The reaction mechanisms of NRR on the B@C2N and B2@C2N are both following an enzymatic pathway, and B2@C2N is a more efficient electrocatalyst with extremely low overpotential of 0.19 eV comparing to B@C2N (0.29 eV). In the low energy region, the hydrogenation of N2 is thermodynamically more favorable than the hydrogen production, thereby improving the selectivity for NRR. Based on these results, a new double-atom strategy may help guiding the experimental synthesis of highly efficient NRR electrocatalysts.