Abstract

Safe autonomous driving critically depends on how well the ego-vehicle can predict the trajectories of neighboring vehicles. To this end, several trajectory prediction algorithms have been presented in the existing literature. Many of these approaches output a multi-modal distribution of obstacle trajectories instead of a single deterministic prediction to account for the underlying uncertainty. However, existing planners cannot handle the multi-modality based on just sample-level information of the predictions. With this motivation, this paper proposes a trajectory optimizer that can leverage the distributional aspects of the prediction in a computationally tractable and sample-efficient manner. Our optimizer can work with arbitrarily complex distributions and thus can be used with output distribution represented as a deep neural network. The core of our approach is built on embedding distribution in Reproducing Kernel Hilbert Space (RKHS), which we leverage in two ways. First, we propose an RKHS embedding approach to select probable samples from the obstacle trajectory distribution. Second, we rephrase chance-constrained optimization as distribution matching in RKHS and propose a novel sampling-based optimizer for its solution. We validate our approach with hand-crafted and neural network-based predictors trained on real-world datasets and show improvement over the existing stochastic optimization approaches in safety metrics.

Abstract

In this article, a novel adaptive controller is designed for Euler-Lagrangian systems under predefined time-varying state constraints. The proposed controller could achieve this objective without a priori knowledge of system parameters and, crucially, of state-dependent uncertainties. The closed-loop stability is verified using the Lyapunov method, while the overall efficacy of the proposed scheme is verified using a simulated robotic arm compared to the state of the art.

Abstract

Discovering new reaction pathways lies at the heart of drug discovery and chemical experimentation. A huge amount of drug reaction data lies in unannotated patent texts which are not machine readable. Reaction roles play an important role in analysing chemical pathways, and tracing chemicals through them, and while there is a vast body of chemical data available, the unavailability of reaction role annotated data is a blocker to effectively deploy deep learning methods for reaction discovery. This paper introduces a new dataset, WEAVE 2.0, obtained from chemical patents, along with full, manual, annotations of novel chemical reactions with reaction role information. We also provide baseline and state of the art models for chemical entity recognition from our raw dataset. Our dataset and associated models form the foundation of neural understanding of chemical reaction pathways via reaction roles.

Abstract

The RSU(Road-Side Unit)-assisted computation offloading allows vehicles to use the RSU’s compute resources to provide connected services which cannot be done due to on-board resource limitations. We propose the empirical analysis framework that can systematically characterize such RSU’s compute workloads. We first formalize the relationship of the three aspects: the local load (generated from vehicles), the composite load (imposed on RSUs) and the traffic flow (mobility patterns of vehicles). Then, our framework takes the models of the local load and the traffic flow as input, and produces the RSUs’ composite loads as an output. We provide the quantitative analysis to show how the RSU’s composite loads change in varying traffic flows in some areas of New York City with different offload patterns.

Abstract

This paper presents a method for automatically generating Wikipedia articles in the Hindi language, using Wikidata as a knowledge base. Our method extracts structured information from Wikidata, such as the names of entities, their properties, and their relationships, and then uses this information to generate natural language text that conforms to a set of templates designed for the domain of interest. We evaluate our method by generating articles about scientists, and we compare the resulting articles to machine-translated articles. Our results show that more than 70% of the generated articles using our method are better in terms of coherence, structure, and readability. Our approach has the potential to significantly reduce the time and effort required to create Wikipedia articles in Hindi and could be extended to other languages and domains as well.

Abstract

Is creativity domain-specific or domain-general? Significant studies have approached this problem using self-reported personality and domain-specific and general creative ideation surveys and lack cognitive perspectives. Creativity is defined as the generation of original and useful products that have emerged through the applications of basic cognitive processes to existing knowledge structures. The current study investigates its implications while modifying the basic knowledge structure. Further, we examine the role of the knowledge domain in creative task performance. For prior knowledge structure, we used advanced Raven’s matrices and modified its constraints to create a novel creative advanced Raven’s Matrices (cAPM) to induce creativity. For creativity, we asked participants to perform two tasks: create Raven's like matrices using creative reasoning task (CRT) and Guilford's alternative uses task (AUT). Unlike other creativity tasks like AUT, CRT allows to measure both convergent and divergent thinking in a given task. The CRT and AUT are different from APM puzzles in two ways, i.e., the knowledge domain and the involvement of creative thinking. The fixed order of APM followed by creativity tasks allowed us to evaluate the boundaries of induced creative thinking across the knowledge domain

Abstract

Accurate localization of sound sources is essential in many acoustic sensing and monitoring applications. In the absence of temporal continuity models, many methods produce unrealistic direction of arrival (DOA) estimates involving sudden changes. To address this, we propose an approach that trains a neural network to predict DOA derivatives in Cartesian coordi- nates (x′, y′, z′), which capture the rate of change in DOA (x, y, z) over time. By combining the predicted DOAs with the predicted derivatives, our method can suppress sudden DOA changes and generate smooth motion trajectories. We introduce an update rule that combines the predicted DOAs with the predicted derivatives to obtain the final DOAs. We validate our approach using the TAU-NIGENS Spatial Sound Events (TNSSE) 2021 dataset. Our results demonstrate that incorporating DOA derivatives improves the accuracy of DOA estimation, particularly in low signal-to- noise ratio scenarios

Abstract

Modern transceiver integrated circuits used in wireless communication systems seek low power wide frequency range receiver front-ends (RxFE) to support multiple communication bands. For this, RxFE leveraging passive N-path mixer first approach are becoming progressively popular due to their low power nature. However, N-path RxFE requires N precise phases of a local oscillator (LO) to control mixer operation, which becomes challenging as the frequency of operation increases (> 10 GHz). In this paper, we give useful insights and present design considerations for the faithful generation of different non- overlapping phases of LO to achieve wide band RF input match. We also present the design and post-layout simulation results of a 4-path mixer based 8-20 GHz RxFE with programmable gain of 10 to 29 dB in TSMC 65 nm CMOS technology. In post-layout simulations, the proposed design achieves an IIP3 of > -10 dB, S11 of < -12 dB and consumes total power of 66 mW from 1.2 V supply.

Abstract

Efficient and reliable non-invasive methods for micro-organism detection have significant implications in various fields such as medical diagnostics, environmental monitoring, and food safety. Existing detection techniques, such as culture-based methods, polymerase chain reaction (PCR) have limitations including time-consuming processes, expensive equipment requirements and invasive sampling procedures. These drawbacks emphasize the need for alternate detection techniques that are non-invasive, rapid, sensitive, specific, and user-friendly. For this purpose, in this research, we present a portable integrated system for micro-organism detection, which is based on electrochemical impedance spectroscopy (EIS). The proposed system includes Zinc Oxide (ZnO) nanorods based interdigitized electrode (IDE) sensor and an EIS circuit. This work also presents the considerations and methodology for the fabrication of ZnO nanorods with high-yield. To validate the efficacy of the proposed methods, experimental results through scanning electron microscope are also presented. Finally, measurement results for detection of a micro-organism (yeast) is also presented to demonstrate the utility and effectiveness of the proposed methods and system.

Abstract

Generative AI has seen remarkable growth over the past few years, with diffusion models being state-of-the-art for image generation. This study investigates the use of diffusion models in generating artificial data generation for electronic circuits for enhancing the accuracy of subsequent machine learning models in tasks such as performance assessment, design, and testing when training data is usually known to be very limited. We utilize simulations in the HSPICE design environment with 22nm CMOS technology nodes to obtain representative real training data for our proposed diffusion model. Our results demonstrate the close resemblance of synthetic data using diffusion model to real data. We validate the quality of generated data, and demonstrate that data augmentation certainly effective in predictive analysis of VLSI design for digital circuits.