检索结果-内蒙古大学图书馆

Opportunistic Query Execution on SmartNICs for Analyzing In-Transit Data

学校读者我要写书评

暂无评论

Opportunistic Query Execution on SmartNICs for Analyzing In-...

IEEE Conference on High Performance Extreme Computing (HPEC)

作者： Jianshen Liu Carlos Maltzahn Craig Ulmer Computer Science & Engineering UC Santa Cruz Santa Cruz CA USA Scalable Modeling & Analysis Sandia National Laboratories Livermore CA USA

High-performance computing (HPC) systems researchers have proposed using current, programmable network interface cards (or SmartNICs) to offload data management services that would otherwise consume host processor cycles in a platform. While this work has successfully mapped data pipelines to a collection of SmartNICs, users require a flexible means of inspecting in-transit data to assess the live state of the system. In this paper, we explore SmartNIC-driven opportunistic query execution, i.e., enabling the SmartNIC to make a decision about whether to execute a query operation locally (i.e., “offload”) or defer execution to the client (i.e., “push-back”). Characterizations of different parts of the end-to-end query path allow the decision engine to make complexity predictions that would not be feasible by the client alone.

关键词：

Graph Laplacian Learning with Exponential Family Noise

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Shi, Changhao Mishne, Gal The Electrical and Computer Engineering Department UC San Diego CA92093 United States The Halıcıoglu Data Science Institute The Neurosciences Graduate Program UC San Diego CA92093 United States

Graph signal processing (GSP) is a prominent framework for analyzing signals on non-Euclidean domains. The graph Fourier transform (GFT) uses the combinatorial graph Laplacian matrix to reveal the spectral decomposition of signals in the graph frequency domain. However, a common challenge in applying GSP methods is that in many scenarios the underlying graph of a system is unknown. A solution in such cases is to construct the unobserved graph from available data, which is commonly referred to as graph or network inference. Although different graph inference methods exist, these are restricted to learning from either smooth graph signals or simple additive Gaussian noise. Other types of noisy data, such as discrete counts or binary digits, are rather common in real-world applications, yet are underexplored in graph inference. In this paper, we propose a versatile graph inference framework for learning from graph signals corrupted by exponential family noise. Our framework generalizes previous methods from continuous smooth graph signals to various data types. We propose an alternating algorithm that jointly estimates the graph Laplacian and the unobserved smooth representation from the noisy signals. We also extend our approach to a variational form to account for the inherent stochasticity of the latent smooth representation. Finally, since real-world graph signals are frequently non-independent and temporally correlated, we further adapt our original setting to a time-vertex formulation. We demonstrate on synthetic and real-world data that our new algorithms outperform competing Laplacian estimation methods that suffer from noise model mismatch. Copyright © 2023, The Authors. All rights reserved.

关键词： Signal processing

Deep learning framework for action prediction reveals multi-timescale locomotor control

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Wei-Chen De Comite, Antoine Voloshina, Alexandra Daley, Monica Seethapathi, Nidhi Department of Electrical Engineering and Computer Science School of Engineering MIT CambridgeMA02139 United States McGovern Institute for Brain Research MIT CambridgeMA02139 United States Brain and Cognitive Sciences School of Science MIT CambridgeMA02139 United States Mechanical and Aerospace Engineering Henry Samueli School of Engineering UC Irvine IrvineCA92697 United States Ecology and Evolutionary Biology School of Biological Sciences UC Irvine IrvineCA92697 United States Biomedical Engineering Henry Samueli School of Engineering UC Irvine IrvineCA92697 United States

Modeling human movement in real-world tasks is a fundamental goal for motor control, biomechanics, and rehabilitation engineering. However, existing models of essential tasks like locomotion are not applicable across varying terrain, mechanical conditions, and sensory contexts. This is at least in part due to simplifying assumptions like linear and fixed timescales mappings between inputs and future actions, which may not be broadly applicable. Here, we develop a deep learning-based framework for action prediction, outperforming traditional models across multiple contexts (walking and running, treadmill and overground, varying terrains) and input modalities (multiple body states, visual gaze). We find that neural network architectures with flexible input history-dependence, like GRU and Transformer, and with architecture-dependent trial embeddings perform best overall. By quantifying the model’s predictions relative to an autoregressive baseline, we identify context- and modality-dependent timescales. These analyses reveal that there is greater reliance on fast-timescale predictions in complex terrain, gaze predicts future foot placement before body states, and the full-body state predictions precede those by center-of-mass states. This deep learning framework for human action prediction provides quantifiable insights into the control of real-world locomotion and can be extended to other actions, contexts, and populations. © 2025, CC BY-NC-SA.

关键词： Deep learning

Optimal Transfer Learning for Missing Not-at-Random Matrix Completion

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Jalan, Akhil Jedra, Yassir Mazumdar, Arya Mukherjee, Soumendu Sundar Sarkar, Purnamrita UT Austin United States MIT United States UC San Diego United States Indian Statistical Institute Kolkata India Department of Computer Science LIDS United States Halıcıoğlu Data Science Institute Department of Computer Science and Engineering United States Department of Statistics and Data Sciences

We study transfer learning for matrix completion in a Missing Not-at-Random (MNAR) setting that is motivated by biological problems. The target matrix Q has entire rows and columns missing, making estimation impossible without side information. To address this, we use a noisy and incomplete source matrix P, which relates to Q via a feature shift in latent space. We consider both the active and passive sampling of rows and columns. We establish minimax lower bounds for entrywise estimation error in each setting. Our computationally efficient estimation framework achieves this lower bound for the active setting, which leverages the source data to query the most informative rows and columns of Q. This avoids the need for incoherence assumptions required for rate optimality in the passive sampling setting. We demonstrate the effectiveness of our approach through comparisons with existing algorithms on real-world biological datasets. © 2025, CC BY.

关键词： Contrastive Learning

Efficient Arm Motion Planning with Reduced Viapoints for Mars Rover Operations

学校读者我要写书评

暂无评论

Efficient Arm Motion Planning with Reduced Viapoints for Mar...

Space Robotics (iSpaRo), International Conference on

作者： Justin Huang Valen Yamamoto Ana Warner Jet Propulsion Laboratory California Institute of Technology Pasadena CA Department of Electrical Engineering and Computer Science UC Irvine Irvine CA Department of Robotics University of Michigan Ann Arbor MI

ISBN: (数字)9798350367232

ISBN: (纸本)9798350367249

This paper describes additions and enhancements we made to the standard RRT-Connect motion planning algorithm to suit the unique needs of operating the robotic arm on the Perseverance Mars Rover. One constraint we have is to maintain a minimum distance between the terrain and any part of the arm hardware. Another goal specific to rover planning is to minimize the number of viapoints in our motion plans. These algorithms should also run efficiently so that rover planners can generate arm motion sequences in time for uplink windows. We applied the RRT-Connect algorithm to the rover’s robotic arm and tested its performance, and we found that it could be used successfully for rover planning with a few changes. First, we implemented the terrain clearance requirement using a software tool called Proximeter. Second, we simplify the paths computed by RRT-Connect to minimize the number of viapoints in a motion plan. Third, we leverage Proximeter to reduce the number of expensive terrain collision checks done by the algorithm, speeding up motion planning.

关键词： Space vehicles Mars Costs Software algorithms Manipulators Hardware Planning

Contrastive Example-Based Control

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Hatch, Kyle Eysenbach, Benjamin Rafailov, Rafael Yu, Tianhe Salakhutdinov, Ruslan Levine, Sergey Finn, Chelsea Department of Computer Science Stanford University United States Machine Learning Department Carnegie Mellon University United States Department of Electrical Engineering and Computer Sciences UC Berkeley United States

While many real-world problems that might benefit from reinforcement learning, these problems rarely fit into the MDP mold: interacting with the environment is often expensive and specifying reward functions is challenging. Motivated by these challenges, prior work has developed data-driven approaches that learn entirely from samples from the transition dynamics and examples of high-return states. These methods typically learn a reward function from high-return states, use that reward function to label the transitions, and then apply an offline RL algorithm to these transitions. While these methods can achieve good results on many tasks, they can be complex, often requiring regularization and temporal difference updates. In this paper, we propose a method for offline, example-based control that learns an implicit model of multi-step transitions, rather than a reward function. We show that this implicit model can represent the Q-values for the example-based control problem. Across a range of state-based and image-based offline control tasks, our method outperforms baselines that use learned reward functions;additional experiments demonstrate improved robustness and scaling with dataset size.1 © 2023, CC BY.

关键词： Reinforcement learning

Diode-Triode Current Mirror Inverter PUF: A Novel Mixed-Signal Low Power Analog PUF

学校读者我要写书评

暂无评论

Diode-Triode Current Mirror Inverter PUF: A Novel Mixed-Sign...

Midwest Symposium on Circuits and Systems (MWSCAS)

作者： C. G. Gisha Ananda Sankar Chakraborty Rajat Subhra Chakraborty Bijoy Antony Jose Jimson Mathew Department of Electronics Cochin University of Science and Technology Kochi India Berkeley Wireless Research Center UC Berkeley Berkeley USA Department of Computer Science and Engineering IIT Kharagpur India Department of Computer Science Cochin University of Science and Technology Kochi India Department of Computer Science Indian Institute of Technology Patna Bihar India

Physically Unclonable Function (PUF) is a useful and versatile lightweight hardware security primitive that takes advantage of unavoidable and unpredictable random process variations during chip manufacturing. This paper proposes a novel diode-triode low-power analog PUF circuit on CMOS current mirror inverter circuit. Due to the random process variation of the MOSFET parameters, the output voltage level of the inverter will vary based on the applied input. These analog voltage variations in the circuit output are converted into a corresponding digital response through an XOR-tree. The performance metrics of the proposed PUF circuit were estimated by simulating a 64-bit PUF circuit at a 32-nano m CMOS technology node. The proposed PUF design achieves a uniformity of 49.87%, an inter-Hamming Distance of 47.52%, average bitaliasing of 49.87%, while consuming 20.99 μW power during operation. The proposed PUF also shows excellent robustness to machine learning based modeling attacks.

关键词：

Developing Human-Centered Intelligent Learning Systems: the application of CARAIX framework 25

学校读者我要写书评

暂无评论

Developing Human-Centered Intelligent Learning Systems: the ...

Adjunct Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization

作者： Miguel Portaz Angeles Manjarrés Olga C. Santos Raúl Cabestrero Pilar Quirós Mar Hermosilla David Puertas-Ramirez Jesus G. Boticario Gadea Lucas Pérez Ana Serrano-Mamolar Álvar Arnaiz-González Miguel Arevalillo-Herráez David Arnau Pablo Arnau-González Raúl Fernández-Matellán David Martin Gomez PhyUM Research Center UNED Madrid Spain PhyUM Research Center UNED Seville Spain Artificial Intelligence UNED Madrid Spain Computer Engineering UBU Burgos Spain Computer Science UV Valencia Spain Systems Engineering and Automation UC3M Madrid Spain

Deploying Traffic Smoothing Cruise Controllers Learned from Trajectory Data

学校读者我要写书评

暂无评论

Deploying Traffic Smoothing Cruise Controllers Learned from ...

IEEE International Conference on Robotics and Automation (ICRA)

作者： Nathan Lichtlé Eugene Vinitsky Matthew Nice Benjamin Seibold Dan Work Alexandre M. Bayen Department of Computer Science ENS Paris-Saclay Paris-Saclay University Department of Mechanical Engineering UC Berkeley Department of Civil and Environmental Engineering Vanderbilt University Department of Mathematics Temple University Department of Physics Temple University Department of Electrical Engineering and Computer Science at UC Berkeley Institute for Transportation Studies UC Berkeley

ISBN: (数字)9781728196817

ISBN: (纸本)9781728196824

Autonomous vehicle-based traffic smoothing con-trollers are often not transferred to real-world use due to challenges in calibrating many-agent traffic simulators. We show a pipeline to sidestep such calibration issues by collecting trajectory data and learning controllers directly from trajectory data that are then deployed zero-shot onto the highway. We construct a dataset of 772.3 kilometers of recorded drives on the I–24. We then construct a simple simulator using the recorded drives as the lead vehicle in front of a simulated platoon consisting of one autonomous vehicle and five human followers. Using policy-gradient methods with an asymmetric critic to learn the controller, we show that we are able to improve average MPG by 11% in simulation on congested trajectories. We deploy this controller to a mixed platoon of 4 autonomous Toyota RAV-4's and 7 human drivers in a validation experiment and demonstrate that the expected time-gap of the controller is maintained in the real world test. Finally, we release the driving dataset [1], the simulator, and the trained controller at https://***/nathanlct/trajectory-training-icra.

关键词： Road transportation Smoothing methods Automation Pipelines Lead Trajectory Calibration