There’s no doubt that social media is a microcosm of information, that covers a great swath of topics. Tweets and touts relaying public policy, to messages that have been forwarded many times spreading misinformation regarding something as serious as COVID-19. The landscape of social media is fertile for the growth and spread of both correct and incorrect information. However, healthcare is a serious issue and laying down a misinformation mine can prove fatal in some cases. Therefore, researchers at SBASSE have created Baang – a digital platform where you use your voice to access reliable healthcare advice!

But why develop a voice-based service? Over the last two decades, voice-based social media platforms have been enabling people who are poor, remote, and low-literate to still get the benefits of the Internet. These platforms allow users to call toll-free phone numbers to record voice messages in their local language and listen to and react to messages recorded by others. Mobile Vaani connects over five million people to infotainment in the media-dark regions in India and CGNet Swara enables rural communities to listen to local news and report grievances. The paper outlines the need for creating a new, versatile application experience. Many scholars have examined the vital role that mainstream social media plays during crises and disasters, for example, by establishing rapid and direct communication channels from authorities, providing support and information to people in need, and bringing to light the challenges on the ground. However, there is a scarcity of research on how voice-based social media platforms are used during public health emergencies by users who are predominantly low-literate and low-income.

To fill this gap, the research team examined three strategies to foster engagement with and dissemination of trusted information: (1) encourage users to access a curated list of approved health guidelines, (2) provide them incentives to engage with and propagate trusted COVID content, and (3) prompt them to reflect on their COVID-related information behaviors. Over a six-month deployment, the platform received around half a million calls from 12,000 users, who were predominately low-literate, low-income men from across Pakistan, with 96% having less than ten years of education. These users recorded over 35,000 audio posts, played them over 2.4 million times, voted on them 322,000 times, and shared them with other users over 130,000 times. The paper also mentions that users approached with all three strategies showed comparatively higher engagement with authentic COVID information. The engagement was not limited to messages being played by the users and included users recording their own COVID content and engaging with and sharing credible content widely with their peers. Users adapted the platform to meet their specific informational, emotional, and instrumental needs.

_{Pictured above is a schematic for the user interface of Baang.}

The findings highlight knowledge engagement as being more meaningful and relevant for information campaigns compared to user engagement with the platform and its features. Our work provides critical insights on how social media platforms can foster user engagement with credible content and, in doing so, makes two important contributions: (1) A six-month deployment of a voice-based social media platform in Pakistan, providing insights into how communities with low literacy rates engaged with credible health information during the COVID pandemic, (2) A mixed-methods analysis that evaluated the efficacy of three design strategies to foster user engagement with health information.

The hope is that through Baang, a major population from the global south, and especially the underserved communities within the Indo-Pak region can get reliable access to health information, navigating safely around the abyss of misinformation found in abundance across social media.

Reference

Agha Ali Raza, Mustafa Naseem, Namoos Hayat Qasmi, Shan Randhawa, Fizzah Malik, Behzad Taimur, Sacha St-Onge Ahmad, Sarojini Hirshleifer, Arman Rezaee and Aditya Vashistha, Fostering Engagement of Underserved Communities with Credible Health Information on Social Media. In Web4Good special track at TheWebConf 2022 (WWW'22), April 25 - 29, 2022, Lyon, France.

Dependency can lead to complacency. The craving to extract oil is partly derived from its incessant need in our industrial sector. Most of the moving parts involved post industrial revolution need some form of fuel, extracted from oil. The cries of scientists from the 1980s fell on partly deaf ears, when they warned of oil scarcity, but the short-term economic benefit of ‘black gold’ was to decapitate our vision for the long term.

Many years later, the question remains still: what will be the alternate when the world runs out of oil? Fortunately, scientists who foresaw this predicament began to work on solutions that may help steer our direction to a more sustainable, renewable resource. One of the products of this enterprise were the birth of electric vehicles (EVs). These machines with futuristic, cyber-punk aesthetics come in all sizes and shapes and appeal to a wide variety of the demography. Some claim them to be a game-changer for the transport industry. Today, electric cars and their charging ports can be seen more frequently in big cities like Lahore, Karachi, and Islamabad. However, like all emerging technologies, electric vehicles come with their own challenges.

Needless to say, that an electric vehicle run on batteries that store charge. With developments in battery technology, efforts have borne fruit to make them more power and cost effective. Still, the high costs of EVs make it a less attractive competitor to traditional gasoline-run cars. Moreover, the lifespan of the batteries installed in EVs are limited, which adds to the already high cost of producing (and owning!) one. Additionally, the reliability of these installed batteries needs to be improved to make charging efficient. Afterall, they compete with a most environmentally unfriendly entity - oil! It is at this junction of reliability on battery technology, and the efforts to understand their health deterioration, that we meet science superheroes from within SBASSE.

Mr. Huzaifa Rauf from the Department of Electrical Engineering, along with Dr. Naveed Arshad from the Department of Computer Science, at SBASSE, are utilizing machine learning to understand and possibly help reduce these problems around EV batteries. The team’s research is focused on lithium-ion batteries, since they are the most advanced and much-sought batteries used in EVs. Mr. Rauf’s team has used multiple machine learning methods to model parameters to understand how a battery degrades. State-of-health and remaining-useful-life are the most important parameters determining a battery’s reliability and lifespan. The team’s primary research objective is determining which machine learning method best models the battery degradation model of lithium-ion batteries. Hence, they presented battery degradation model in several machine learning methods and algorithms. The results of these models were presented to compare their accuracy, and ability to handle the complex data.

These models can also reduce the strain on laboratory-based methods for understanding battery health degradation and produce results faster. This research presented by Mr. Rauf’s team is still at its early stages. However, such tools are elemental to understand battery health and reliability, which will enable the vast use of electric vehicles.

Their most recent paper highlighting this work was published in the journal Renewable and Sustainable Energy Reviews, a 15.0 impact factor journal. We congratulate them on such a tremendous achievement and wish them best of fortune in their future work.

Reference

Huzaifa Rauf, Muhammad Khalid, Naveed Arshad, Machine learning in state of health and remaining useful life estimation: Theoretical and technological development in battery degradation modelling, Renewable and Sustainable Energy Reviews, Volume 156, 2022, 111903, ISSN 1364-0321, https://doi.org/10.1016/j.rser.2021.111903.