Sentence Embeddings and High-Speed Similarity Search for Fast Computer Assisted Annotation of Legal Documents

Human-performed annotation of sentences in legal documents is an important prerequisite to many machine learning based systems supporting legal tasks. Typically, the annotation is done sequentially, sentence by sentence, which is often time consuming and, hence, expensive. In this paper, we introduce a proof-of-concept system for annotating sentences “laterally.” The approach is based on the observation that sentences that are similar in meaning often have the same label in terms of a particular type system. We use this observation in allowing annotators to quickly view and annotate sentences that are semantically similar to a given sentence, across an entire corpus of documents. Here, we present the interface of the system and empirically evaluate the approach. The experiments show that lateral annotation has the potential to make the annotation process quicker and more consistent.

Protein-Coding Genes in Euarchontoglires with Pseudogene Homologs in Humans

An original bioinformatics technique is developed to identify the protein-coding genes in rodents, lagomorphs and nonhuman primates that are pseudogenized in humans. The method is based on per-gene verification of local synteny, similarity of exon-intronic structures and orthology in a set of genomes. It is applicable to any genome set, even with the number of genomes exceeding 100, and efficiently implemented using fast computer software. Only 50 evolutionary recent human pseudogenes were predicted. Their functional homologs in model species are often associated with the immune system or digestion and mainly express in the testes. According to current evidence, knockout of most of these genes leads to an abnormal phenotype. Some genes were pseudogenized or lost independently in human and nonhuman hominoids.

A Fast Computer-Generated Holographic Method for VR and AR Near-Eye 3D Display

A fast computer-generated holographic method with multiple projection images for a near-eye VR (Virtual Reality) and AR (Augmented Reality) 3D display is proposed. A 3D object located near the holographic plane is projected onto a projection plane to obtain a plurality of projected images with different angles. The hologram is calculated by superposition of projected images convolution with corresponding point spread functions (PSF). Holographic 3D display systems with LED as illumination, 4f optical filtering system and lens as eyepiece for near-eye VR display and holographic optical element (HOE) as combiner for near-eye AR display are designed and developed. The results show that the proposed calculation method is about 38 times faster than the conventional point cloud method and the display system is compact and flexible enough to produce speckle noise-free high-quality VR and AR 3D images with efficient focus and defocus capabilities.

Rapid Simulation of Flat Knitting Loops Based On the Yarn Texture and Loop Geometrical Model

In order to create realistic loop primitives suitable for the fast computer-aided design (CAD) of the flat knitted fabric, we have a research on the geometric model of the loop as well as the variation of the loop surface. Establish the texture variation model based on the changing process from the normal yarn to loop that provides the realistic texture of the simulative loop. Then optimize the simulative loop based on illumination variation. This paper develops the computer program with the optimization algorithm and achieves the loop simulation of different yarns to verify the feasibility of the proposed algorithm. Our work provides a fast CAD of the flat knitted fabric with loop simulation, and it is not only more realistic but also material adjustable. Meanwhile it also provides theoretical value for the flat knitted fabric computer simulation.