Leveraging Automatic High-Level Synthesis Resource Sharing to Maximize Dynamical Voltage Overscaling with Error Control

Approximate Computing has emerged as an alternative way to further reduce the power consumption of integrated circuits (ICs) by trading off errors at the output with simpler, more efficient logic. So far the main approaches in approximate computing have been to simplify the hardware circuit by pruning the circuit until the maximum error threshold is met. One of the critical issues, though, is the training data used to prune the circuit. The output error can significantly exceed the maximum error if the final workload does not match the training data. Thus, most previous work typically assumes that training data matches with the workload data distribution. In this work, we present a method that dynamically overscales the supply voltage based on different workload distribution at runtime. This allows to adaptively select the supply voltage that leads to the largest power savings while ensuring that the error will never exceed the maximum error threshold. This approach also allows restoring of the original error-free circuit if no matching workload distribution is found. The proposed method also leverages the ability of High-Level Synthesis (HLS) to automatically generate circuits with different properties by setting different synthesis constraints to maximize the available timing slack and, hence, maximize the power savings. Experimental results show that our proposed method works very well, saving on average 47.08% of power as compared to the exact output circuit and 20.25% more than a traditional approximation method.

FPGA-Implemented Fractal Decoder with Forward Error Correction in Short-Reach Optical Interconnects

Forward error correction (FEC) codes combined with high-order modulator formats, i.e., coded modulation (CM), are essential in optical communication networks to achieve highly efficient and reliable communication. The task of providing additional error control in the design of CM systems with high-performance requirements remains urgent. As an additional control of CM systems, we propose to use indivisible error detection codes based on a positional number system. In this work, we evaluated the indivisible code using the average probability method (APM) for the binary symmetric channel (BSC), which has the simplicity, versatility and reliability of the estimate, which is close to reality. The APM allows for evaluation and compares indivisible codes according to parameters of correct transmission, and detectable and undetectable errors. Indivisible codes allow for the end-to-end (E2E) control of the transmission and processing of information in digital systems and design devices with a regular structure and high speed. This study researched a fractal decoder device for additional error control, implemented in field-programmable gate array (FPGA) software with FEC for short-reach optical interconnects with multilevel pulse amplitude (PAM-M) modulated with Gray code mapping. Indivisible codes with natural redundancy require far fewer hardware costs to develop and implement encoding and decoding devices with a sufficiently high error detection efficiency. We achieved a reduction in hardware costs for a fractal decoder by using the fractal property of the indivisible code from 10% to 30% for different n while receiving the reciprocal of the golden ratio.

Distribution-free, Risk-controlling Prediction Sets

While improving prediction accuracy has been the focus of machine learning in recent years, this alone does not suffice for reliable decision-making. Deploying learning systems in consequential settings also requires calibrating and communicating the uncertainty of predictions. To convey instance-wise uncertainty for prediction tasks, we show how to generate set-valued predictions from a black-box predictor that controls the expected loss on future test points at a user-specified level. Our approach provides explicit finite-sample guarantees for any dataset by using a holdout set to calibrate the size of the prediction sets. This framework enables simple, distribution-free, rigorous error control for many tasks, and we demonstrate it in five large-scale machine learning problems: (1) classification problems where some mistakes are more costly than others; (2) multi-label classification, where each observation has multiple associated labels; (3) classification problems where the labels have a hierarchical structure; (4) image segmentation, where we wish to predict a set of pixels containing an object of interest; and (5) protein structure prediction. Last, we discuss extensions to uncertainty quantification for ranking, metric learning, and distributionally robust learning.

Link-Layer Retransmission-based Error-Control Protocols in FSO Communications: A Survey

Free-space optical (FSO) communications have gained significant interest over the last few years, thanks to the capability to transport extremely high-speed data over long distances without exhausting radio frequency (RF) resources. FSO communication is widely considered in various network scenarios, such as inter-satellite/deep-space links, ground-station/vehicles, satellite/aerial links, or terrestrial links. It is expected to be a key enabling technology for the next generation of 6G wireless networks. Nevertheless, despite the great potential of FSO communications, its performance suffers from various limitations and challenges: atmospheric turbulence, clouds, weather conditions, and pointing misalignment. The error-control solutions, including physical layer (PHY) and link-layer methods, aim to mitigate the transmission errors caused by such adverse issues. While the existing surveys on error-control solutions in FSO systems primarily focussed on the PHY methods, we instead provide a review of link-layer solutions. In particular, we conduct an extensive literature survey of state-of-the-art retransmission protocols, both automatic repeat request (ARQ) and hybrid ARQ (HARQ), for various FSO communication scenarios, including point-to-point terrestrial, cooperative, multi-hop relaying, hybrid FSO/RF, satellite/aerial, and deep-space systems. Furthermore, we provide a survey of recent literature and insightful discussion on the cross-layer design frameworks related to link-layer retransmission protocols in FSO communication networks. Finally, the lessons learned, design guidelines, related open issues, and future research directions are exposed.

Link-Layer Retransmission-based Error-Control Protocols in FSO Communications: A Survey

Free-space optical (FSO) communications have gained significant interest over the last few years, thanks to the capability to transport extremely high-speed data over long distances without exhausting radio frequency (RF) resources. FSO communication is widely considered in various network scenarios, such as inter-satellite/deep-space links, ground-station/vehicles, satellite/aerial links, or terrestrial links. It is expected to be a key enabling technology for the next generation of 6G wireless networks. Nevertheless, despite the great potential of FSO communications, its performance suffers from various limitations and challenges: atmospheric turbulence, clouds, weather conditions, and pointing misalignment. The error-control solutions, including physical layer (PHY) and link-layer methods, aim to mitigate the transmission errors caused by such adverse issues. While the existing surveys on error-control solutions in FSO systems primarily focussed on the PHY methods, we instead provide a review of link-layer solutions. In particular, we conduct an extensive literature survey of state-of-the-art retransmission protocols, both automatic repeat request (ARQ) and hybrid ARQ (HARQ), for various FSO communication scenarios, including point-to-point terrestrial, cooperative, multi-hop relaying, hybrid FSO/RF, satellite/aerial, and deep-space systems. Furthermore, we provide a survey of recent literature and insightful discussion on the cross-layer design frameworks related to link-layer retransmission protocols in FSO communication networks. Finally, the lessons learned, design guidelines, related open issues, and future research directions are exposed.