Search

Domain Generalization (DG) is essentially a sub-branch of out-of-distribution generalization, which trains models from multiple source domains and generalizes to unseen target domains. Recently, some domain generalization algorithms have emerged, but most of them were designed with non-transferable complex architecture. Additionally, contrastive learning has become a promising solution for simplicity and efficiency in DG. However, existing contrastive learning neglected domain shifts that caused severe model confusions. In this paper, we propose a Dual-Contrastive Learning (DCL) module on feature and prototype contrast. Moreover, we design a novel Causal Fusion Attention (CFA) module to fuse diverse views of a single image to attain prototype. Furthermore, we introduce a Similarity-based Hard-pair Mining (SHM) strategy to leverage information on diversity shift. Extensive experiments show that our method outperforms state-of-the-art algorithms on three DG datasets. The proposed algorithm can also serve as a plug-and-play module without usage of domain labels., Inadequate proof of the effectiveness of the method

Recently, out-of-distribution (OOD) generalization has attracted attention to the robustness and generalization ability of deep learning based models, and accordingly, many strategies have been made to address different aspects related to this issue. However, most existing algorithms for OOD generalization are complicated and specifically designed for certain dataset. To alleviate this problem, nicochallenge-2022 provides NICO++, a large-scale dataset with diverse context information. In this paper, based on systematic analysis of different schemes on NICO++ dataset, we propose a simple but effective learning framework via coupling bag of tricks, including multi-objective framework design, data augmentations, training and inference strategies. Our algorithm is memory-efficient and easily-equipped, without complicated modules and does not require for large pre-trained models. It achieves an excellent performance with Top-1 accuracy of 88.16% on public test set and 75.65% on private test set, and ranks 1st in domain generalization task of nicochallenge-2022.

Serious environmental pollution will occur if corn stalk materials are discarded randomly or burned in open air. This study presents a possible method for using corn stalk materials to make...

The properties of asphalt binder modified with corn stalk fiber were investigated in this study. The physical-mechanical behaviors of the fiber modified asphalt binder were studied by the penetration and softening point tests, dynamic shear test, bending beam rheometer test and rotational viscosity test. The results indicate that the addition of corn stalk fiber decreases the penetration and increases the softening point of the asphalt binder. Furthermore, the addition of the corn stalk fiber increases the complex modulus and decreases the phase angle of the asphalt. The creep stiffness of the asphalt binder at low temperatures first decreases and then increases with an increase in corn stalk fiber content. The variation of the m-value of the asphalt binder is the opposite to which of the creep stiffness. The corn stalk fiber has good uniformity in the asphalt binder, followed by basalt fiber and finally lignin fiber. The viscosity results indicate that the addition of corn stalk fiber can increase the viscosity of the asphalt binder, and the temperature sensitivity of the asphalt binder decreases.

The chemical and rheological properties of polymer modified bitumen incorporating bio-oil derived from waste cooking oil (WCO) were investigated in this paper. At first, the chemical composition and mixing mechanism of the experimental materials were analysed from the perspective of functional group, and the influence of bio-oil on the activation energy was also researched. Then, the effect of bio-oil on the rotational viscosities of polymer modified bitumen and construction temperatures of corresponding mixtures was studied. Finally, the shear and bending rheological properties of polymer modified bitumen containing bio-oil were investigated. The results show that the bio-oil and styrene–butadiene–styrene (SBS) modified bitumen is mainly physically mixed, the addition of bio-oil decreases the activation energy of SBS modified bitumen. Additionally, the SBS modified bitumen containing bio-oil has lower viscosity values, and corresponding mixtures also have lower construction temperatures. Furthermore, the addition of bio-oil in SBS modified bitumen reduces the shear modulus and increases the bending creep compliance, which means bio-oil has positive effect on the low-temperature thermal cracking resistance performance while sacrificing the high-temperature rutting resistance performance to some extent. Therefore, the incorporation of WCO-based bio-oil in polymer modified bitumen is a promising technique to be used in cold regions where the low-temperature problems are more crucial.

In this study, a bioresource material, corn stalks, were converted by a combination of physical and chemical methods into a corn stalk fiber material that can be utilized in an asphalt pavement. Firstly, corn stalk fiber was produced with different comminution times, mass fractions of sodium hydroxide solution, reaction times, and reaction temperatures. An orthogonal experimental method was employed to determine the optimal production process. The results show that the optimal comminution time was 3.5 min, and corn stalk should be reacted with a sodium hydroxide solution at 80 &deg, C for 30 min to obtain the best corn stalk fiber product. The mass ratios of corn stalk fibers, sodium hydroxide, and water should be 8:1:200. After an evaluation of the physical and mechanical properties of the corn stalk fibers, asphalt binders, and mixtures, it was concluded that this kind of corn stalk fiber can be a good substitute of pavement fiber.