Search

We investigate the task of retrieving information from compositional distributed representations formed by hyperdimensional computing/vector symbolic architectures and present novel techniques that achieve new information rate bounds. First, we provide an overview of the decoding techniques that can be used to approach the retrieval task. The techniques are categorized into four groups. We then evaluate the considered techniques in several settings that involve, for example, inclusion of external noise and storage elements with reduced precision. In particular, we find that the decoding techniques from the sparse coding and compressed sensing literature (rarely used for hyperdimensional computing/vector symbolic architectures) are also well suited for decoding information from the compositional distributed representations. Combining these decoding techniques with interference cancellation ideas from communications improves previously reported bounds (Hersche et al., 2021) of the information rate of the distributed representations from 1.20 to 1.40 bits per dimension for smaller codebooks and from 0.60 to 1.26 bits per dimension for larger codebooks. [ABSTRACT FROM AUTHOR]

Internet of Things (IoT) paradigm broadens to several research fields. Thus, Wireless Body Area Network (WBAN) has been adopted as a standard to create an IoT scheme implemented on healthcare system. Furthermore, IoT can be employed to measure several diseases including stroke, diabetes, as well as respiratory diseases to monitor patient condition and environment change that might be harmful for patients. Implementation and realization of IoT for monitoring respiratory diseases is needed since in Taiwan the risk to get more severe symptoms by those diseases is relatively high considering the air pollution that is getting higher. In this study, IoT system is built based on the integration of several independent applications. In addition, our scheme consists of four main components such as environment sensing box, patient monitoring tools, android apps, and web Graphical User Interface (GUI). Web GUI is useful for health practitioners such as doctors and nurses to monitor the condition of patients obtained by patient monitoring tools. Moreover, patients and doctors can assess the status of weather and environmental condition whether it is safe or harmful for patients. Finally, android based apps is necessary for patients to connect all of schemes as well as monitoring all conditions including health and environmental status. The assessment of our tools indicates that the implemented scheme is consistent enough shown by low Root Mean Square Error (RMSE) and Mean Average Percentage Error (MAPE) achieved by our IoT system. [ABSTRACT FROM AUTHOR]

This brief presents a novel low-power analog Euclidian vector normalizer for analog signal processing in vector forms, such as the preprocessing of sensor array signals. Instead of using multiple stages of translinear circuits to synthesize the Euclidian norm with square root, squared sum, and division functions, this circuit performs concurrent Euclidian vector normalization on input signals with a feedback control loop to achieve a compact area and low power consumption. A simple two-step calibration is also provided to overcome device mismatches. This 6-channel normalizer is part of an analog gas sensing front end and was implemented in a 65-nm low-power CMOS technology with an area of 0.053 mm2. It consumes a maximum current of $2.08~{\mu }\text{A}$ from 0.6/0.65-V supplies with a response time of 6.5 ${\mu }\text{s}$ , thus achieving a low energy per operation of 1.39 pJ/channel. [ABSTRACT FROM AUTHOR]

The field of machine learning is witnessing rapid advances along several fronts: new machine learning models, new machine learning algorithms utilizing these models, new hardware architectures for these algorithms, and new technologies for creating energy-efficient implementations of such hardware architectures. Hyperdimensional (HD) computing represents one such model. Emerging nanotechnologies, such as carbon nanotube field-effect transistors (CNFETs), resistive random-access memory (RRAM), and their monolithic 3D integration, enable energy- and area-efficient hardware implementations of HD computing architectures. Such efficient implementations are achieved by exploiting several characteristics of the component nanotechnologies (e.g., energy-efficient logic circuits, dense memory, and incrementers naturally enabled by gradual reset of RRAM cells) and their monolithic 3D integration (enabling tight integration of logic and memory), as well as various characteristics of the HD computing model (e.g., embracing randomness that allows us to utilize rather than avoid inherent variations in RRAM and CNFETs, resilience to errors in the underlying hardware). We experimentally demonstrate and characterize an end-to-end HD computing nanosystem built using monolithic 3D integration of CNFETs and RRAM. Using our nanosystem, we experimentally demonstrate the pairwise classification of 21 languages with measured mean accuracy of up to 98% on >20 000 sentences (6.4 million characters), training using one text sample (~100 000 characters) per language, and resilient operation (98% accuracy) despite 78% of bits in HD representation being stuck at 0 or 1 in hardware. We also show that the monolithic 3D implementations of HD computing can have 35 $\times $ improved energy-execution time product for training and inference of language classification data sets (while using 3 $\times $ less area) compared to silicon CMOS implementations. [ABSTRACT FROM AUTHOR]