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.
The proliferation of unmanned vehicles offers many opportunities for solving environmental sampling tasks with applications in resource monitoring and precision agriculture. Informative path planning (IPP) includes a family of methods which offer improvements over traditional surveying techniques for suggesting locations for observation collection. In this work, we present a novel solution to the IPP problem by using a coregionalized Gaussian processes to estimate a dynamic scalar field that varies in space and time. Our method improves previous approaches by using a composite kernel accounting for spatiotemporal correlations and at the same time, can be readily incorporated in existing IPP algorithms. Through extensive simulations, we show that our novel modeling approach leads to more accurate estimations when compared with formerly proposed methods that do not account for the temporal dimension.
We aim to provide a general framework of for computational photography that recovers the real scene from imperfect images, via the Deep Nonparametric Convexified Filtering (DNCF). It is consists of a nonparametric deep network to resemble the physical equations behind the image formation, such as denoising, super-resolution, inpainting, and flash. DNCF has no parameterization dependent on training data, therefore has a strong generalization and robustness to adversarial image manipulation. During inference, we also encourage the network parameters to be nonnegative and create a bi-convex function on the input and parameters, and this adapts to second-order optimization algorithms with insufficient running time, having 10X acceleration over Deep Image Prior. With these tools, we empirically verify its capability to defend image classification deep networks against adversary attack algorithms in real-time.
Small devices are frequently used in IoT and smart-city applications to perform periodic dedicated tasks with soft deadlines. This work focuses on developing methods to derive efficient power-management methods for periodic tasks on small devices. We first study the limitations of the existing Linux built-in methods used in small devices. We illustrate three typical workload/system patterns that are challenging to manage with Linux's built-in solutions. We develop a reinforcement-learning-based technique with temporal encoding to derive an effective DVFS governor even with the presence of the three system patterns. The derived governor uses only one performance counter, the same as the built-in Linux mechanism, and does not require an explicit task model for the workload. We implemented a prototype system on the Nvidia Jetson Nano Board and experimented with it with six applications, including two self-designed and four benchmark applications. Under different deadline constraints, our approach can quickly derive a DVFS governor that can adapt to performance requirements and outperform the built-in Linux approach in energy saving. On Mibench workloads, with performance slack ranging from 0.04 s to 0.4 s, the proposed method can save 3% - 11% more energy compared to Ondemand. AudioReg and FaceReg applications tested have 5%- 14% energy-saving improvement. We have open-sourced the implementation of our in-kernel quantized neural network engine. The codebase can be found at: https://github.com/coladog/tinyagent.
Existing volumetric methods for predicting 3D human pose estimation are accurate, but computationally expensive and optimized for single time-step prediction. We present TEMPO, an efficient multi-view pose estimation model that learns a robust spatiotemporal representation, improving pose accuracy while also tracking and forecasting human pose. We significantly reduce computation compared to the state-of-the-art by recurrently computing per-person 2D pose features, fusing both spatial and temporal information into a single representation. In doing so, our model is able to use spatiotemporal context to predict more accurate human poses without sacrificing efficiency. We further use this representation to track human poses over time as well as predict future poses. Finally, we demonstrate that our model is able to generalize across datasets without scene-specific fine-tuning. TEMPO achieves 10$\%$ better MPJPE with a 33$\times$ improvement in FPS compared to TesseTrack on the challenging CMU Panoptic Studio dataset.
We introduce an interpretable deep learning approach for direction of arrival (DOA) estimation with a single snapshot. Classical subspace-based methods like MUSIC and ESPRIT use spatial smoothing on uniform linear arrays for single snapshot DOA estimation but face drawbacks in reduced array aperture and inapplicability to sparse arrays. Single-snapshot methods such as compressive sensing and iterative adaptation approach (IAA) encounter challenges with high computational costs and slow convergence, hampering real-time use. Recent deep learning DOA methods offer promising accuracy and speed. However, the practical deployment of deep networks is hindered by their black-box nature. To address this, we propose a deep-MPDR network translating minimum power distortionless response (MPDR)-type beamformer into deep learning, enhancing generalization and efficiency. Comprehensive experiments conducted using both simulated and real-world datasets substantiate its dominance in terms of inference time and accuracy in comparison to conventional methods. Moreover, it excels in terms of efficiency, generalizability, and interpretability when contrasted with other deep learning DOA estimation networks.
Digital twins have attracted a great deal of recent attention from a wide range of fields. A basic requirement for digital twins of nonlinear dynamical systems is the ability to generate the system evolution and predict potentially catastrophic emergent behaviors so as to providing early warnings. The digital twin can then be used for system "health" monitoring in real time and for predictive problem solving. In particular, if the digital twin forecasts a possible system collapse in the future due to parameter drifting as caused by environmental changes or perturbations, an optimal control strategy can be devised and executed as early intervention to prevent the collapse. Two approaches exist for constructing digital twins of nonlinear dynamical systems: sparse optimization and machine learning. The basics of these two approaches are described and their advantages and caveats are discussed.
Benefiting from tens of GHz of bandwidth, terahertz (THz) communications has become a promising technology for future 6G networks. However, the conventional hybrid beamforming architecture based on frequency-independent phase-shifters is not able to cope with the beam split effect (BSE) in THz massive multiple-input multiple-output (MIMO) systems. Despite some work introducing the frequency-dependent phase shifts via the time delay network to mitigate the beam splitting in THz wideband communications, the corresponding issue in reconfigurable intelligent surface (RIS)-aided communications has not been well investigated. In this paper, the BSE in THz massive MIMO is quantified by analyzing the array gain loss. A new beamforming architecture has been proposed to mitigate this effect under RIS-aided communications scenarios. Simulations are performed to evaluate the effectiveness of the proposed system architecture in combating the array gain loss.
To alleviate the shortage of spectral resources as well as to reduce the weight, volume, and power consumption of wireless systems, joint communication-radar (JCR) systems have become a focus of interest in both civil and military fields. JCR systems based on time-modulated arrays (TMAs) constitute an attractive solution as a benefit of their high degree of beam steering freedom, low cost, and high accuracy. However, their sideband radiation results in energy loss, which is an inherent drawback. Hence the energy-efficiency optimization of TMA-based JCR systems is of salient importance, but most of the existing TMA energy-efficiency optimization methods do not apply to JCR systems. To circumvent their problems, a single-sideband structure is designed for flexibly reconfigurable energy-efficient TMA beam steering. First, some preliminaries on single-sideband TMAs are introduced. Then, a closed-form expression is derived for characterizing the energy efficiency. Finally, the theoretical results are validated by simulations.
We present an approach for computationally efficient dynamic time warping (DTW) and clustering of time-series data. The method frames the dynamic warping of time series datasets as an optimisation problem solved using dynamic programming, and then clusters time series data by solving a second optimisation problem using mixed-integer programming (MIP). There is also an option to use k-medoids clustering for increased speed, when a certificate for global optimality is not essential. The improved efficiency of our approach is due to task-level parallelisation of the clustering alongside DTW. Our approach was tested using the UCR Time Series Archive, and was found to be, on average, 33% faster than the next fastest option when using the same clustering method. This increases to 64% faster when considering only larger datasets (with more than 1000 time series). The MIP clustering is most effective on small numbers of longer time series, because the DTW computation is faster than other approaches, but the clustering problem becomes increasingly computationally expensive as the number of time series to be clustered increases.