This work studies the semantic segmentation of 3D LiDAR data in dynamic scenes for autonomous driving applications. A system of semantic segmentation using 3D LiDAR data, including range image segmentation, sample generation, inter-frame data association, track-level annotation and semisupervised learning, is developed. To reduce the considerable requirement of fine annotations, a CNN-based classifier is trained by considering both supervised samples with manually labeled object classes and pairwise constraints, where a data sample is composed of a segment as the foreground and neighborhood points as the background. A special loss function is designed to account for both annotations and constraints, where the constraint data are encouraged to be assigned to the same semantic class. A dataset containing 1838 frames of LiDAR data, 39934 pairwise constraints and 57927 human annotations is developed. The performance of the method is examined extensively. Qualitative and quantitative experiments show that the combination of a few annotations and large amount of constraint data significantly enhances the effectiveness and scene adaptability, resulting in greater than 10% improvement.

Aberrant programmed cell death protein 1 (PD-1) expression on the surface of T cells is known to inhibit T cell effector activity and to play a pivotal role in tumor immune escape; thus, maintaining an appropriate level of PD-1 expression is of great significance. We identified KLHL22, an adaptor of the Cul3-based E3 ligase, as a major PD-1–associated protein that mediates the degradation of PD-1 before its transport to the cell surface. KLHL22 deficiency leads to overaccumulation of PD-1, which represses the antitumor response of T cells and promotes tumor progression. Importantly, KLHL22 was markedly decreased in tumor-infiltrating T cells from colorectal cancer patients. Meanwhile, treatment with 5-fluorouracil (5-FU) could increase PD-1 expression by inhibiting the transcription of KLHL22. These findings reveal that KLHL22 plays a crucial role in preventing excessive T cell suppression by maintaining PD-1 expression homeostasis and suggest the therapeutic potential of 5-FU in combination with anti–PD-1 in colorectal cancer patients.

The effects of the endophytic fungus Gilmaniella sp. and its elicitor on the defense and metabolic responses of host plants Atractylodes lancea were investigated, in order to understand how to utilize endophytic fungi and their elicitor resources better. The results showed that the promotion effect of the fungus on the growth of host plantlets was much better than that of its elicitor. Both fungus and elicitor enhanced defense-related enzyme activities. In fungus-inoculated groups, phenylalanine ammonia lyase and polyphenol oxidase activities increased slowly, and reached a maximum level during the later stages, whereas peroxidase activity peaked in the first few days. Additionally, the activities of chitinase and β-1,3-glucanase were significantly higher than those of the control plants. In elicitor-treated groups, however, most of the enzymes were activated during the early stage, and their highest levels were generally lower than those of the fungus-inoculated groups. Compared with the elicitor, fungal infection improved the photosynthetic rate of the host, and increased carbohydrate levels as well as chlorophyll content in host leaves. The total content of the four main components of volatile oil was also increased in elicitor-treated groups, but there was no particular pattern in this increase. Meanwhile, in the fungus-inoculated groups, the content of atractylone significantly increased with time, while the content of β-eudesmol decreased. These results indicated that fungal elicitor could substantially improve the total content of volatile oil, while the fungus could more effectively enhance the quality of herbal medicines.

With the explosive development of information technology, vulnerabilities have become one of the major threats to computer security. Most vulnerabilities with similar patterns can be detected effectively by static analysis methods. However, some vulnerable and non-vulnerable code is hardly distinguishable, resulting in low detection accuracy. In this paper, we define the accurate identification of vulnerabilities in similar code as a fine-grained vulnerability detection problem. We propose VulSniper which is designed to detect fine-grained vulnerabilities more effectively. In VulSniper, attention mechanism is used to capture the critical features of the vulnerabilities. Especially, we use bottom-up and top-down structures to learn the attention weights of different areas of the program. Moreover, in order to fully extract the semantic features of the program, we generate the code property graph, design a 144-dimensional vector to describe the relation between the nodes, and finally encode the program as a feature tensor. VulSniper achieves F1-scores of 80.6% and 73.3% on the two benchmark datasets, the SARD Buffer Error dataset and the SARD Resource Management Error dataset respectively, which are significantly higher than those of the state-of-the-art methods.