Computing attention is the backbone of transformer-based models like large language models. However, the increasing diversity of attention algorithms presents significant challenges for unleashing hardware performance. State-of-the-art variants like FlashAttention target a specific attention algorithm or hardware platform, which fail to generalize to other algorithms and platforms.

Modern Deep Neural Network (DNN) exhibits a pattern where multiple sub-models are executed, guided by control flows such as loops and switch/merge operations. This dynamic nature introduces complexities in batching the requests of such DNNs for efficient execution on GPUs. In this paper, we present NeuStream, a programming model and runtime system for serving deep learning workloads using stream processing. NeuStream decomposes the inference workflow into modules and forms them into a streaming processing system where a request flows through. Based on such abstraction, NeuStream is able to batch requests at fine-grained module granularity. To maximize serving goodput, NeuStream exploits a two-level scheduling approach to decide the best batching requests and resource allocation for each module while satisfying service level objectives (SLOs). Our evaluation of NeuStream on a set of modern DNNs like Large Language Models (LLM) and diffusion models, etc., shows that NeuStream significantly improves goodput compared to state-of-the-art DNN serving systems.

The sparse Mixture-of-Experts (MoE) architecture is increasingly favored for scaling Large Language Models (LLMs) efficiently, but it depends on heterogeneous compute and memory resources. These factors jointly affect system Cost, Accuracy, and Performance (CAP), making trade-offs inevitable. Existing benchmarks often fail to capture these trade-offs accurately, complicating practical deployment decisions. To address this, we introduce MoE-CAP, a benchmark specifically designed for MoE systems. Our analysis reveals that achieving an optimal balance across CAP is difficult with current hardware; MoE systems typically optimize two of the three dimensions at the expense of the third—a dynamic we term the MoE-CAP trade-off. To visualize this, we propose the CAP Radar Diagram. We further introduce sparsity-aware performance metrics—Sparse Memory Bandwidth Utilization (S-MBU) and Sparse Model FLOPS Utilization (S-MFU)—to enable accurate performance benchmarking of MoE systems across diverse hardware platforms and deployment scenarios. This benchmark is available on Github: https://github.com/sparse-generative-ai/MoE-CAP.

As AI chips incorporate numerous parallelized cores to scale deep learning (DL) computing, inter-core communication is enabled recently by employing high-bandwidth and low-latency interconnect links on the chip (e.g., Graphcore IPU). It allows each core to directly access the fast scratchpad memory in other cores, which enables new parallel computing paradigms. However, without proper support for the scalable inter-core connections in current DL compilers, it is hard for developers to exploit the benefits of this new architecture.

Embedding-based retrieval methods construct vector indices to search for document representations that are most similar to the query representations. They are widely used in document retrieval due to low latency and decent recall performance. Recent research indicates that deep retrieval solutions offer better model quality, but are hindered by unacceptable serving latency and the inability to support document updates. In this paper, we aim to enhance the vector index with end-to-end deep generative models, leveraging the differentiable advantages of deep retrieval models while maintaining desirable serving efficiency. We propose Model-enhanced Vector Index (MEVI), a differentiable model-enhanced index empowered by a twin-tower representation model. MEVI leverages a Residual Quantization (RQ) codebook to bridge the sequence-to-sequence deep retrieval and embedding-based models. To substantially reduce the inference time, instead of decoding the unique document ids in long sequential steps, we first generate some semantic virtual cluster ids of candidate documents in a small number of steps, and then leverage the well-adapted embedding vectors to further perform a fine-grained search for the relevant documents in the candidate virtual clusters. We empirically show that our model achieves better performance on the commonly used academic benchmarks MSMARCO Passage and Natural Questions, with comparable serving latency to dense retrieval solutions.

Current state-of-the-art document retrieval solutions mainly follow an index-retrieve paradigm, where the index is hard to be directly optimized for the final retrieval target. In this paper, we aim to show that an end-to-end deep neural network unifying training and indexing stages can significantly improve the recall performance of traditional methods. To this end, we propose Neural Corpus Indexer (NCI), a sequence-to-sequence network that generates relevant document identifiers directly for a designated query. To optimize the recall performance of NCI, we invent a prefix-aware weight-adaptive decoder architecture, and leverage tailored techniques including query generation, semantic document identifiers, and consistency-based regularization. Empirical studies demonstrated the superiority of NCI on two commonly used academic benchmarks, achieving +21.4% and +16.8% relative enhancement for Recall@1 on NQ320k dataset and R-Precision on TriviaQA dataset, respectively, compared to the best baseline method.

Water-body segmentation is an important issue in remote sensing and image interpretation. Classic methods for counteracting this problem usually include the construction of index features by combining different spectra, however, these methods are essentially rule-based and fail to take advantage of context information. Additionally, as the quality of image resolution improves, these methods are proved to be inadequate. With the rise of convolutional neural networks (CNN), the level of research about segmentation has taken a huge leap, but the field is still facing an increasing demand for data and the problem of blurring boundaries. In this letter, a new segmentation network called restricted receptive field deconvolution network (RRF DeconvNet) is proposed, with which to extract water bodies from high-resolution remote sensing images. Compared with natural images, remote sensing images have a weaker pixel neighborhood relativity; in consideration of this challenge, an RRF DeconvNet compresses the redundant layers in the original DeconvNet and no longer relies on a pretrained model. In addition, to tackle the blurring boundaries that occur in CNN, a new loss function called edges weighting loss is proposed to train segmentation networks, which has been shown to significantly sharpen the segmentation boundaries in results. Experiments, based on Google Earth images for water-body segmentation, are presented in this letter to prove our method.