Sign Language Production (SLP) is the process of converting the complex input text into a real video. Most previous works focused on the Text2Gloss, Gloss2Pose, Pose2Vid stages, and some concentrated on Prompt2Gloss and Text2Avatar stages. However, this field has made slow progress due to the inaccuracy of text conversion, pose generation, and the rendering of poses into real human videos in these stages, resulting in gradually accumulating errors. Therefore, in this paper, we streamline the traditional redundant structure, simplify and optimize the task objective, and design a new sign language generative model called Stable Signer. It redefines the SLP task as a hierarchical generation end-to-end task that only includes text understanding (Prompt2Gloss, Text2Gloss) and Pose2Vid, and executes text understanding through our proposed new Sign Language Understanding Linker called SLUL, and generates hand gestures through the named SLP-MoE hand gesture rendering expert block to end-to-end generate high-quality and multi-style sign language videos. SLUL is trained using the newly developed Semantic-Aware Gloss Masking Loss (SAGM Loss). Its performance has improved by 48.6% compared to the current SOTA generation methods.

In this paper, we propose a dual-condition diffusion pre-training model named SignDIFF that can generate human sign language speakers from a skeleton pose. SignDiff has a novel Frame Reinforcement Network called FR-Net, similar to dense human pose estimation work, which enhances the correspondence between text lexical symbols and sign language dense pose frames, reduces the occurrence of multiple fingers in the diffusion model. In addition, we propose a new method for American Sign Language Production (ASLP), which can generate ASL skeletal pose videos from text input, integrating two new improved modules and a new loss function to improve the accuracy and quality of sign language skeletal posture and enhance the ability of the model to train on largescale data. We propose a simple baseline for ASL production and report the scores of 17.19 and 12.85 on BLEU-4 on the How2Sign dev/test sets. We evaluated our model on the previous mainstream dataset PHOENIX14T, and our method achieved the SOTA results. In addition, our image quality far exceeds all previous results by 10 percentage points in terms of SSIM.

Deletion based view maintenance is a building block in many query explanation and analytics applications, such as sensitivity analysis, what-if analysis, data cleaning, and probabilistic databases, which require exploring how a query's output result changes when excluding (deleting) subsets of input relations (referred to as “deletion interventions”). These applications' ability to remain interactive is limited by how fast they can recompute output results, especially over complex queries and large data. A critical limitation in existing engines is their performance. To remain interactive, existing approaches either restrict the types of aggregation functions, require explanations to be pre-specified, or apply sampling and approximation. We show how to build an interactive deletion intervention evaluation engine for SPJA queries, by Ieveraging recent advances in fast fine-grained provenance that captures input/output relationship for each physical operator in a denormalized compact pointer-free format. We use provenance to propagate the deletion status of tuples per operator, in a tight loop that leads to improvement in instruction and data locality. In addition, this representation is amenable to a variety of physical optimizations, such as dictionary-encoding, bit-packing, vectorization, and parallelization. Operators' output tuples can be evenly split across threads, and their deletion status can be computed independently. Our preliminary results show that on the TPC-H workload, our engine takes on average 9ms to evaluate 1K deletion interventions at a time, achieving on average wins of 3 orders of magnitude over DBToaster, a state-of-the art IVM based engine, and wins of 4 orders of magnitude against existing provenance based approaches. Utilizing tight loops and a cache-friendly provenance and intervention representation enable us to evaluate thousands of deletion interventions at interactive speed, and enable interactive time query explanations over more complex queries and data than previously possible.

Stream Window Join (SWJ), a vital operation in stream analytics, struggles with achieving a balance between accuracy and latency due to out-of-order data arrivals. Existing methods predominantly rely on adaptive buffering, but often fall short in performance, thereby constraining practical applications. We introduce PECJ, a solution that proactively incorporates unobserved data to enhance accuracy while reducing latency, thus requiring robust predictive modeling of stream oscillation. At the heart of PECJ lies a mathematical formulation of the posterior distribution approximation (PDA) problem using variational inference (VI). This approach circumvents error propagation while meeting the low-latency demands of SWJ. We detail the implementation of PECJ, striking a balance between complexity and generality, and discuss both analytical and learning-based approaches. Experimental evaluations reveal PECJ's superior performance. The successful integration of PECJ into a multi-threaded SWJ benchmark testbed further establishes its practical value, demonstrating promising advancements in enhancing data stream processing capabilities amidst out-of-order data.

What-if queries are the building blocks for many explanation and analytics applications—sensitivity analysis, hypothetical reasoning, data cleaning, probabilistic databases—that explore how a query's output changes due to input data changes. Their response time is bounded by intervention evaluation latency, which can be in the minute or hours for complex queries and large datasets. FaDE is a compilation engine that uses provenance to evaluate hypothetical deletion and scaling interventions at low latency and high throughput. FaDE forgoes conventional provenance representations as symbolic expressions and leverages their underlying relational structure. This accelerates intervention evaluation on average by 1000× against IVM and 10,000× against prior provenance-based approaches. In addition, FaDE develops a suite of optimizations (e.g., compilation, parallelization, incremental evaluation, sparse representations) that collectively raise evaluation throughput to >1 million interventions per sec—a rate that can brute-force existing applications within 1 s.

BACKGROUND: Whole exome sequencing (WES) has been widely used in human genetics research. BGISEQ-500 is a recently established next-generation sequencing platform. However, the performance of BGISEQ-500 on WES is not well studied. In this study, we evaluated the performance of BGISEQ-500 on WES by side-to-side comparison with Hiseq4000, on well-characterized human sample NA12878. RESULTS: BGISEQ demonstrated similarly high reproducibility as Hiseq for variation detection. Also, the SNVs from BGISEQ data is highly consistent with Hiseq results (concordance 96.5%~ 97%). Variation detection accuracy was subsequently evaluated with data from the genome in a bottle project as the benchmark. Both platforms showed similar sensitivity and precision in SNV detection. While in indel detection, BGISEQ showed slightly higher sensitivity and lower precision. The impact of sequence depth and read length on variation detection accuracy was further analyzed, and showed that variation detection sensitivity still increasing when the sequence depth is larger than 100x, and the impact of read length is minor when using 100x data. CONCLUSIONS: This study suggested that BGISEQ-500 is a qualified sequencing platform for WES.