Two-step wireless link quality prediction using multi-camera images

Given the recent advances in wireless communication technology, all things are being connected to networks, and it is expected that various new and novel applications will be created. The diversification in applications will yield various requirements for wireless communication. We believe that autonomous robot services will be popular after COVID, and the management, monitoring, and operation of the mobility robots requires highly reliable wireless links. In this paper, we propose a wireless link quality prediction system that uses camera images. The proposal generates a prediction model for each camera and combines the output of the models using weights calculated by the outputs of reliability models. By providing separate prediction models for each camera, the prediction system easily handles new additional cameras and drops the cameras found to have low reliability. Indoor experiments show that the proposed prediction scheme outperforms the prediction method that uses all camera images as input features without regard to their reliability. © 2022 IEEE.

Efficient Clustering of Genetic Entities

The paper proposes a novel application of a highly efficient method for comparing symbol sequences based on convolution. The technique utilizes Fast Fourier Transform (FFT) to compare long symbol sequences achieving practical results using commodity PC hardware. While the main focus is on bioinformatics, the proposed approach is general and can work beyond genetic sequences. One of the main advantages of the proposed method is the robustness to insertion/deletion. Also, unlike standard alignment algorithms, the proposed method is parameter-free. The paper shows that the FFT-based comparison allows for efficient clustering of long sequences in bioinformatics as a practical application. Exploration of coronaviruses offers an illustration of the proposed clustering techniques. © 2022 IEEE.

Virtual Safety Assistant: An Efficient Tool for Ensuring Safety During Covid-19 Pandemic

Using intelligent virtual assistants for controlling employee population in workspaces is a research area that remains unexplored. This paper presents a novel application of virtual humans to enforce Covid-19 safety measures in a corporate workplace. For this purpose, we develop a virtual assistant platform, Chloe, equipped with automatic temperature sensing, facial recognition, and dedicated chatbots to act as an initial filter for ensuring public health. Whilst providing an engaging user interaction experience, Chloe minimizes human to human contact, thus reducing the spread of infectious diseases. Chloe restricts the employee population within the office to government-approved safety norms. We experimented with Chloe as a virtual safety assistant in a company, where she interacted and screened the employees for Covid-19 symptoms. Participants filled an online survey to quantify Chloe’s performance in terms of interactivity, system latency, engagement, and accuracy, for which we received positive feedback. We performed statistical analysis on the survey results that reveal positive results and show effectiveness of Chloe in such applications. We detail system architecture, results and limitations. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Detecting Emerging Associations and Behaviors with Regional and Diachronic Word Embeddings

We introduce novel applications of the word embedding association test (WEAT)-a method for assessing differential biases and attitudes in word embeddings-for identifying correlations of human attitudes and behaviors with word embedding associations, and for automatically detecting words associated with a concept. We assess our methods by measuring the evolution of associations related to COVID-19, using survey data from the COVID States project as validation, along with a set of COVID-19 validation words developed based on surveys and sample responses created by expert psychologists studying COVID-19 behavior. We first show that word associations measured using the WEAT correlate with the behaviors and attitudes of the population which produced an embedding's training corpus. We take Pearson's \rho between word embedding associations from a diachronic set of English-language word embeddings with COVID States survey data related to COVID-19 attitudes and behaviors. We find statistically significant correlations between WEAT associations and survey results for 19 of 23 survey questions, with Pearson's \rho as high as.96. Survey responses for 10 questions correlate with WEAT associations in embeddings trained on Twitter data from several weeks prior to the survey. We also introduce the unipolar word embedding association test (U-WEAT), which measures strength of association with a single attribute word group, rather than between two opposing polar attribute groups. In an embedding trained on Twitter data from Oregon, the U-WEAT returns a positive effect size for 88% of validation words based on their association with a COVID-19 concept group, where less than 20% of the embedding vocabulary has a positive effect size, despite the prevalence of language related to COVID-19 during the time period in which the corpus was trained. A qualitative analysis of other words identified by the U-WEAT reveals a wide array of people, places, behaviors, and attitudes related to COVID-19. © 2022 IEEE.