SCIENCE NEWS

The Gaia space telescope of European Space Agency (ESA) has released a data set consisting of the speeds and chemical composition of almost 2 billion stars, that gives us a new perspective of our universe. This new perspective has been claimed to have enabled us to look back in time of each stars to deduce the whereabouts and status of each star. The Gaia telescope is located at a distance of approximately 1.5 million kilometers from earth being launched in 2013, has released a set of data comprising of brightness and positions of 1.1 billion stars in 2016 and another set with colours, temperatures and radii of 1.7 billion stars in 2018. The current addition of data now avails us with more information in our quest to understand the origin of stars and our universe. The chemical make up sheds light on where the stars
formed based on the metal they contain, that is obtained through the starlight analyses. The article from New Scientist describes the study findings in detail.

Complete article: New Scientist, June 18, 2022, page 10

The article discusses a latest study where switching off a gene in the liver using a one-off injection could permanently reduce the risk of heart diseases, as observed in mice. This technology is called epigenome-editing where the injection controls the gene activity. Liver cells produce a protein called PCSK9 which break downs another protein that is responsible for the removal of cholesterol from blood. Given the protein that removes cholesterol is preserved by reducing or stopping production of PCSK9 should lead to lowering of cholesterol. Although there are drugs that adopted this method to lower cholesterol in humans, they require periodic-frequent injections. Permanently altering DNA to quell production of PCSK9 could lead to complications and cells becoming cancerous, this new gene-editing methods claimed no observed complications as such. Should this new method be seen promising, could solve the issue of cholesterol related diseases in humans.

Complete article: New Scientist, June 18, 2022, page 16

News emerged as an engineer attached to Google had been suspended, after he had claimed that sentient AI had been developed. This article that appeared in New Scientist (June 18-24, 2022, page 11: No sign of a machine mind yet) has discussed the claim and refuted by Google. Sentient AI refers to the potential of an AI system to mimic human consciousness.

Blake Lemoine, the engineer, had released transcripts of the conversations with the AI he had assisted in developing called LaMDA (Language Model for Dialogue Applications), that had consisted of dialogues where the AI discusses about its feelings of fear, happy and sad based on activities that never have happened. Lemoine claimed that LaMDA can not only make convincing conversations, but can also present itself having self-awareness and feelings. On the other hand, Adrian Weller at the Turing Institute in London claimed, that although LaMDA is an impressive system, it is no-where close to being sentient. The article goes on to discuss if the current trajectory of AI with umpteen amounts of data used to train several models will be successful in creating an artificial mind. In the meantime, Adrian Hilton of University of Surrey, UK claimed that he would not believe that we understood the mechanisms behind what makes something sentient as yet and that what we do in terms of AI research using machine learning is really intelligence. Weller claims, that given human emotions rely entirely on sensory inputs, we may be able to attain sentient systems, but not as yet.

Complete article: New Scientist, June 18, 2022, page 11

The article claims to have observed revival of seven types of viruses that laid dormant for thousands of years in the Siberian permafrost (with the thawing of the frost). The youngest of which have been dormant for 27,000 years while the oldest (revived latest) was in the ice for 48,500 years. Simultaneously, the article reveals that several other scientists have claimed to have revived microorganisms that were dormant in sediments, ice and salt crystals for more than 250 million years, which are yet to be confirmed not to have been contaminated. These ancient viruses have the potential to spread and infect living organisms including humans.

Complete article: New Scientist, December 03, 2022, page 17

Another news item is discussed in this issue, on carbon capture, where the United Kingdom has opened the largest facility to absorb carbon dioxide from the atmosphere to produce sodium bicarbonate. And it was discussed that we would need more innovative solutions and upscaling to see the effect of carbon capture on revival of our climate. This article that appeared on New Scientist discussed about another form of carbon capture. The oceans are currently the biggest sink for carbon dioxide, while the method discussed in this news item is to enhance the capacity of the oceans to absorb carbon by chemically modifying it. This has been given a dangerous warning as this would permanently affect the sensitive marine environments, while it is thought that the lack of progress otherwise in curbing carbon emissions, necessitating such measures. To avoid global warming breaching the 1.5⁰C rise (the world agreed to in 2015 in Paris), it is computed that a removal of 584 billion tonnes of carbon dioxide from the atmosphere between 2020 – 2100 is imperative. This accounts for approximately 7 billion tonnes of carbon dioxide a year. Planting trees to aid the process of removing carbon dioxide is seen as insufficient in our quest. Oceans covering almost 70% of the area, technologies to sequester carbon using the ocean has a potential of removing 100 billion tonnes per year while the constraints such as feasibility shrink that figure to 5 – 10 billion a year.

Other than ocean based growth systems such as kelp forests, seagrass and mangroves, iron fertilization is seen as a potential alternative. Seeding ocean waters with the nutrient iron will increase the population of carbon-absorbing phytoplankton. The organisms down the food-chain will lock carbon dioxide biologically. Another scientist David King proposed mimicking the natural process of whale feces that provide nutrients for the growth of phytoplanktons, the results of a small scale study is awaited.

Supplying nutrients to the ocean surface are also being tried through engineering exercises, that mimic the process of upwelling that naturally occurs in parts of the ocean due to currents. The artificial upwelling was also intended to bring the nutrient rich cool waters from the bottom of the oceans to the surface and sustain phytoplankton growth. However, extensive growth of phytoplankton in ocean waters is believed to deteriorate ocean environments, mainly by reducing the dissolved oxygen levels and oxygen diffusion in ocean environments.

As an alternative, alkalinity enhancement is considered for carbon sequestering process using ocean waters. At the moment, with increased levels of carbon dioxide in the atmosphere, we are making the ocean water acidic due to the formation of carbonic acid as carbon dioxide dissolves in water. Acidification of ocean water reduces its capability to dissolve further carbon dioxide while also imparting detrimental effect on the marine life forms. Alkalinity enhancement is believed to enhance carbon absorption capacity while also reducing the impact acidity would have on marine living organisms. This technology is in the initial stages and is seen a promising alternative for our quest. The article discusses the projects that have used these techniques and the pros and cons each of them harboured.

Another article that appeared on Science (volume 378, page 22 –  23rd December, 2022) discusses this phenomenon, having completed a study of adding alkaline lime powder in Apalachicola Bay off Florida, and found that it drew down carbon while also reducing acidification. This was claimed in the article as the first field study on the technique known as ocean liming, which also ended up as a successful implementation of the technique. The author argues that this is a natural process, and the technique only accelerates the process. This is further reinstated by the report from National Academies of Science, Engineering and Medicine (NASEM) that called for $2.5 billion in researches on similar techniques, more specifically on field studies. 

Complete article 01: New Scientist, July 02, 2022, page 48

Complete article 02: Science, December 23, 2022, page 22

Carbon emissions have been blamed for climate change and its impacts on the environment. While the globe agrees to reduce emissions, it has always been spoken of to absorb carbon dioxide in the atmosphere to assist maintain the global rise in temperatures to 1.5 ⁰C. The United Kingdom inaugurated the largest carbon dioxide absorption project on 24^th July, a factory that produces bicarbonate for dialysis machines, pharmaceutical tablets and baking soda. The news article on New Scientist discusses the project. The project initiated by Tata Chemicals Europe (TCE) plant in Norwhich, UK is expected to absorb 36,000 tonnes of carbon dioxide a year eventually increasing the capacity to capture 40,000. The gas is captured from a gas-fired power plant. Instead of storing the captured gas, it is purified and transformed into liquid to facilitate the production of sodium bicarbonate. The article highlights the incapability of the plant to store the captured gas, where a fraction of it is turned into gas again. It is noted that this is not an ultimate solution to the crisis, but a little step towards managing the gas in the atmosphere.

Complete article: New Scientist, July 02, 2022, page 16

The article talks about time and discusses 9 questions that bogle human perception of time, including what is time?, why does time only go one way?, how do we sense time? What affects our perception of time? How much time does evolution take? can time lead us to a theory of everything? How do we make the most of our time? Can we live without time? Will time ever end?

 Complete article: New Scientist, 18 June, 2022, page 41

The demand for electrical power is on an exponential rise, together increasing the consumption of batteries. This causes two issues; one is the demand that requires unprecedented demand for materials and the other is the management of wasted batteries, which are toxic to the ecosystems. Retrieving materials in possibly elemental form does incur a huge cost in the recovery processes. However, with time this may be seen plausible due to;

1 Unprecedented increase in demand

2 Available of new technologies that make the recovery process economically feasible

3 Unavailability of raw materials, that in turn increase the cost of them

The biggest demand for batteries comes from the increased consumption in electronic vehicles (EVs) that use predominantly lithium, cobalt and nickel. It has been estimated that the demand for Li will be more than 20 times of that today, by 2050. Simultaneously, the waste amounts to 600,000 metric tons of recyclable li-ion batteries is expected to increase to 1.6 million metric tons by 2030. While recycling alone cannot meet the increase in demand, it can certainly supply a significant amount to reduce the stress on mining of raw material. According to the article that appeared on New Scientist, the laboratories and companies are now in the quest pumping money in to materials engineering, looking for a way to produce metals in the purest form, economically. Several companies such as Li-Cycle (Canadian) and American Battery Technology Company have already increased their capacity of recycling in anticipation of increasing recovery of metals. Redwood Materials which recycles li-ion batteries from electronics and EVs, with a new plant in Reno with an investment of $3.5 billion, expects to produce 1 million li-ion EV batteries by 2025. 

Complete article: MIT Technology Review, Jan/Feb, 2023, page 48

The article was based on a study where nine people who had lower body paralysis regained the ability to walk with the assistance of a frame, after receiving a prolonged electrical stimulation to the injured area of their spine. The article claims that the study led to the identification of neurons that can assist the patient improve in the condition post paralysis. Spine, usually working with electrical stimulations received form brain to move the muscles and hence facilitate the ability to walk, has been receiving electrical treatment to relieve pain and in cases to regain maneuverability given sufficient number of neurons remained unscathed in the injured zone. The study discussed in the article was conducted by Bloch and her colleagues, who implanted electrical devices in nine people, six of the patients had some feeling in their legs with no ability to walk while the other three had neither the feelings nor the ability to walk.

The findings form the study: immediately after the first treatment, the patients could walk 25 m with the aid of a frame in 6 minutes. On a continued treatment over 5 months, in addition to physiotherapy 5 times a week, enabled them to walk 50 m in 6 minutes on average. Four of the patients could even walk without any electrical stimulation at the end of the treatment period.

A further study in mice by the same team of scientists by monitoring brain activity during injury and electrical stimulation and subsequent regain of activity enabled them to identify the neurons that are responsible for the ability to walk. These neurons when switched off, completely paralyzed mice, while when they were switched back on, they regained the ability to maneuver. This finding outlines the promise of potential advancement in treatment for paralysis.

Complete article: New Scientist, November 19, 2022, page 25            

The article that appeared on New Scientist discussed on how big a number can get and how to write them figuratively and for computational exercises. For example, consider the number of sand grains on earth? The study of extremely large numbers is called googology. The book, ‘The biggest number in the world: a journey to the edge of mathematics’ by David Darling and Agnijo Banerjee elucidates the googology by starting from normal big numbers such as 10 billion trillion or the number of ways a pack of cards can be arranged (8.0658×10⁶⁷). The book then goes on to talk about the method to write larger values. One such method of writing larger numbers is Knuth’s up-arrow notation, which was named after Donald Knuth.

Knuth’s up-arrow notation: consider multiplication as simply repeated addition: 5×3 = 5+5+5 while exponentiation is repeated multiplication: 5³ = 5×5×5. Up-arrow notation uses a single up-arrow to mean exponentiation, so that 5↑3 indicates 5³. Two arrows then mean repeated  exponentiation: 5↑↑3 = 5↑(5↑5). This pattern continues as such that each new arrow meaning the steps before should be repeated, resulting swiftly in immense numerical power. The number 2↑↑↑↑4, for example is so large that it cannot be displayed in digital form. 

Complete article: New Scientist, June 18, 2022, page 37

The previous record of recovering 1 million year old DNA in 2021 has now been broken with the recovery of a 2 million year old DNA from sediments by Eske Willerslev and team from University of Cambridge. This would assist recreation of the ecosystem that existed in Greenland at that time. This oldest DNA has been recovered from Kobenhavn Formation, which a series of sand, silt and mud laid around 2 million years ago. As this DNA was bound to mineral particles, rather than fossilized organisms, the enzymes couldn’t access and decay it, leaving it preserved. The DNA comprises mostly of plants (some of which are still in Greenland) and some from animals, such as arctic hares, rodents, muskats, reindeer and an elephant relative. This discovery has given hope that even older DNA could be found in the arctic seas, preserved.   

Complete article: New Scientist, December 17, 2022, page 12

Computing one’s carbon footprint to delineate environmental friendliness of people and institutions have become popular among modern humans, who have already started bearing the brunt of impact the industrial revolution had on the environment. No matter how hard we strive to halt pumping carbon dioxide in the atmosphere, we are set to increase the fraction of the gas in the atmosphere for several decades to come. The article that appeared on New Scientist (25 June, 2022, page 11: Food miles matter less than you might think) discusses on another factor that contributes to carbon footprint. Food miles simply indicate the carbon footprint of a food item that has been transported from the location where it was grown to where it was consumed. Essentially, the longer the food travel, greater the food miles and so is greater the carbon footprint. In other words, if one person eats locally, he/she reduces food miles for their consumption, that has inherently less carbon footprint. The article however, analyses the credibility of the claim that lower the food miles, lower the carbon footprint, drawing attention to the carbon footprint of locally produced food items. It was found that the carbon footprint is not as significant as it was considered it to be between food items with different food miles. It is worthy to note the following regarding carbon footprint and food production. The production of the food  we eat is responsible for more than a third of global greenhouse gas emissions and that 20% of food-related emissions are from food miles. It is therefore crucial that consumers resort to measures that would curb carbon emissions from their food consumption practices and growing locally is not a solution.

Complete article: New Scientist, June 25, 2022, page 11

A patient from New Zealand became the first person to receive a gene-editing treatment to permanently lower her cholesterol. CRISPR is the editing tool that was used which has moved from labs to clinic in the last decade. Compared to CRISPR 2.0 (base editing tool), CRISPR 3.0 (prim editing) allows insertion of chunks of DNA into a genome. The article claims that this could possibly be adapted to treat blood pressure and certain other diseases in the future.

Complete article: MIT Technology Review, Jan/Feb, 2023, page 27

Human enhancement: ‘improving human being beyond what is considered typical or normal’. This definition includes prosthetic limbs that outperform biological limbs or drugs that boost cognitive capacities beyond typical human range. We very well know that such enhancements are proscribed in sports and is considered a punishable offense. However, research and development does occur at rapid pace in the science community, which imperatively calls for regularization. European Commission endorsed a set of ethics guidelines to be adopted in human enhancement related research, which is now included in the Horizon Europe funding programme.

The biggest challenge in such ethical guidelines lies in the very definition of the term, human enhancement. The definition entails, modification aimed at improving  human performance—as opposed to restoring it. Restoration of human ability to the natural levels would mean that the human enhancements are applied to physically challenged people, which is not the case discussed in definition. Furthermore, abilities may differ, and the term ‘healthy’ and ‘normal’ may not be applicable generally to everyone. Hence the terminology in ethical guidelines needs negotiations.

Discussed in the article appeared on Science; Cyborg artist and activist Neil Harbisson had an antenna implanted and osseointegrated in his skull which was qualified as human enhancement according to the definition. However, he had the ability to convert colours and images into audible vibration, that included wavelengths of IR and UV, which are not visible to human eye.

Some typical human enhancement research areas listed in the article are prosthetics, tissue engineering, genome editing, neurotechnolog,y and nanomedicine, belonging to field of biomedical engineering. Therapeutic or otherwise, the yearning to enhance ones ability is not new, yet, research and development in the area needs regularization. The question asked in the article is that since human enhancement is not part of medicine, laws and ethics related to medicine may not be applicable. Medical ethics guidelines depend on, well-being, autonomy, informed consent, equality, justice and (moral or social) responsibility. Human enhancements on the other hand are required to include, liberty, bodily integrity and human dignity, which are currently not included.  

Complete article: Science, November 25, 2022, page 835

Artificial intelligence has been used in various applications, trying to mimic human behaviour. Creativity has been considered as a talent and not even available to all humans (not scientifically proven), yet, artificial intelligence is now pouted to get into the creative arts as well. In 2021, DALL-E was released and shocked the world, with its ability to text-to-image model (in addition to several other tools that even write scientific articles, including critical review of literature). In this tool of text-image, it is capable of creating a visual art from any piece of descriptive text. While Google launched its own image-making AI called Imagen, DALL-E2 was unveiled in April/2022 with a massive leap forward, as it claimed. In the meantime, the article claims that the biggest game-changer was Stable Diffusion, an open-source text-to-image from a UK-based startup Stability AI in August.

Just like language models, AI trains models based on available data on the internet, that may have bias. The article elaborates on the challenge that just like language models, text-to-image generators can amplify the biased and toxic associations embedded in the training data. With this technology already transforming in to text-to-video tools, the creativity is highly challenged with randomness. The article claims that nothing else in AI grabbed peoples’ interest more in the last year—for the best and worst reasons.

Complete article: MIT Technology Review, Jan/Feb, 2023, page 48

The article claims that the emission-free cars and trucks will likely make for 13% of all new auto sales globally in 2022, according to International Energy Agency. This trend is expected to surpass 30% by the end of this decade.

Several causes have propelled this trend, specifically the government policies. Auto companies are gearing up for this trend by setting up supply chains, building manufacturing capacity and better and effective designs. The Hongguang Mini, a small EV with a price tag less than $5,000 has become the most preferred in the world. The challenges object the trend is mostly the charging options and battery supply and clean technology in power production. 

Complete article: MIT Technology Review, Jan/Feb, 2023, page 27

Another story on a plant that changes carbon dioxide to bicarbonate has been reported in this issue. This is another plant designed to reduce the CO₂ in to methanol has been opened and running in China. Methanol is used primarily in production of chemicals such as plastics and also used as a fuel and is naturally made from coal or natural gas. Omat Freyr Sigurbjornsson from the Carbon Recycling International (CRI) expects that this plant would curb , tonnes of CO₂ a year from the atmosphere. The estimate includes the emissions from the facility as well.

This facility in the city of Anyang by Shunli captures CO₂ from emissions during the production of lime with hydrogen turns 160,000 tonnes of CO₂ into 110,000 tonnes of methanol per year.  they contain, that is obtained through the starlight analyses. The article from New Scientist describes the study findings in detail. CRI is working on a second CO₂-to-methanol site in China, to recycle  petrochemical complex emission to produce methanol for plastic production.

Complete article: New Scientist, November 12, 2022, page 11

Salt is used on roads in winter, in countries which receive snow-fall which transforms ice into slush. Addition of salt reduces the melting point of water. A similar observation could be made when sea water freezes, it squeezes salt out making water beneath more saltier, which makes lower level water hard to freeze. An experiment showed that ice cubes covered with kitchen salt reached –(negative)10⁰C after a few seconds when ice cubes without salt remained at 0 ⁰C.bLilley Prasher developed a method to make this process reversible, the idea known as ionocaloric refrigeration. This idea takes advantage of the large temperature change—and therefore the large heat absorption—obtained by melting this special form of ice when put in contact with specific salt.

Complete article: Science, December 23, 2022, page 1275

Can AI models be trusted to shape our future?

Many possible cases could be assumed for the static picture of a dynamic process shown in Figure, of which, three are discussed here.

Case 01: the base ball player pitching the ball on top of a bas-ball field

Case 02: a man throwing a baseball at a pitcher on a baseball field

Case 03: a baseball player at bat and a catcher in the dirt during a baseball game

For a human viewing the picture, he would base the conclusive statement he/she would make on the experience they have together with a logical understanding of the context of the picture. Depending on this analogy, the first case would be best depiction of the picture, while the other two would rarely be even discussed. On the other hand, when AI is making a judgement on the picture, it would have been based on rough approximations based on statistical similarities to the images that were used to develop the model, and instilling logic in the judgement is highly questionable. For example, any human would know that there would be other players in the vicinity but not in the image, where AI would find it hard to make a guess on that aspect in making sense of the picture.

AI failing to identify the difference between a man riding a horse and a horse riding a man may be considered inconsequential in terms of a risk potential. However, the article speaks about an incident where an AI driven car failed stop when a man carrying a stop sign was encountered. The AI would emphatically identify a stop sign in the usual locations as seen in the training data or a man, which were distinctively well trained and tested. However, what the AI system failed to respond is when both events happened unexpectedly (not in the training data). For cases where zero tolerance towards failures is aspired, depending on the AI is questionable.

The article identified that the biggest teams of researchers in AI are no longer part of the academic circles and publications where peer review process is essential prior to publication and rigorous scrutiny at different levels is inevitable prior to implementation.

The article identified that the biggest teams of researchers in AI are no longer part of the academic circles and publications where peer review process is essential prior to publication and rigorous scrutiny at different levels is inevitable prior to implementation. In the corporate sectors, conducting researches in the field, bypass the scrutiny process and this is said in the article as, ‘we see what the company wants us to see’. These developments by the corporate are only to be used in controlled environments, given they were developed from fair experiments, which cannot be used in a real-scale application, which is usually known as ‘demoware’ in IT sectors.

The article states that the deep learning techniques have advanced the pattern recognition ability in data analytical methods, yet, have intricately three major flaws, that need attention when implementing.

1. The patterns learnt are ironically, superficial and not conceptual

2. The results it creates are hard to interpret (often said to be black boxes)

3. The results are difficult to use in the context of other processes, such as memory and reasoning (this forbids the general application of the models, strange to the controlled environment they were trained with).

The article quoted Les Valiant, a computer scientist from Harvard university as saying, ‘the central challenge [going forward] is to unify the formulation of … learning and reasoning’. At the mean time it is said in the article, that we have reached the local minimum in which companies pursue benchmarks rather than foundational ideas. What is actually required is to intertwine logical approach and a more conceptual approach in the development of the model to substantiate the learning and reasoning in the resulting models. This reinstates the need for basic research, which depends on such an approach. The article closes with an example. Suppose aliens learn about the human behaviour using such an AI system, that was based on the shadows of humans, what would the replications of the model when implemented. For example, all humans will have the same colour, and some shadows may change in size based on the time (location of sun) and worse of all, the shadows may disappear in nights. The outcome of their model would be uncharacteristic, illogical and inappropriate.

It could be put, the AI is as intelligent as its training data, or it is true as of now. The outcome of AI model when treated as black boxes devoid of fundamental concept, may not be trusted unless it was experimentally verified. An uncertainty will linger always, that as explained in the example with a man carrying a stop sign, may not be trustworthy. As mentioned earlier, specially in applications where zero tolerance is envisaged from a model prediction, the outcome of an AI model may be questionable. The way forward would be to instill logical and fundamental approach in model development.

 Complete article: Scientific American, October, 2022, page 46

Frequency and intensity of drought, melting of glaciers and extreme precipitation events are dynamically changing contemporarily, and how they will affect the future of water cycle depends on calculation of water circulation. The warming earth bring more water from surface of earth to the atmosphere, specifically through evaporation, depleting surface and ground water resources. With the increasing in demand primarily due to population growth further stress the water cycle unprecedently.

Intensification of water’s circulation through atmosphere, land and ground affect the sustainability of water supply. This in turn affect the water cycle through which the frequency and intensity of events are further altered. The process of water distribution and conveyance among atmosphere, land and ground is severely affected by the warming climate that in turn affect agricultural systems, weather events and freshwater resources. The change in the water cycle also varies across climate zone, further stressing the cycle in an unpredictable way. The models predict a drastic change in the extreme events in the near future.

A paper by Thomas G Huntington et al. in 2019 introduced a new framework for quantifying the intensity of the terrestrial water cycle, known as water cycle intensity, is the sum of precipitation (groundwater input) and actual evapotranspiration (groundwater output) over a time period. liberty, bodily integrity and human dignity, which are currently not included. This parameter when analysed together with GIS tools could help understand the spatial distribution of water cycle intensity. Next steps will have to focus on finding links between the anthropogenic factors (such as industry, lifestyle and technology)  and their effects on the water cycle intensification that is occurring contemporarily.

Complete article: American Scientist, Volume 110, page 324

An article appeared on New Scientist (25 June 2022, page 16: How can we prevent AI from being racist, sexist and offensive?)  discusses the potential of AI systems being based on our inherent prejudices. The article begins with a case speaking about the biases in facial recognition working best with white people compared to the black communities. Considering that most AI systems depend on machine learning models that depend on data., the author quotes from Carissa Veliz from University of Oxford as proposing two solutions to wring bias out of AI, 1) vetting the data used to train models and 2) applying filter to intercept any harmful outputs. However, the article highlights the challenges owing to the difficulty in engineering the vast data that already had been collected and that the repercussions may affect the AI systems underperforming in other domains that relied on the raw data, that included the biases. For example, the models that are based on images, captions, texts on websites and other sources, would all inherently include biases, as the human nature inevitably encompasses that in nature. Predicting if a prisoner would repeat the crime would essentially require the biases that it may entail in the raw data collected and tampering or engineering the data would make the models invalid per se. A youtuber Yannic Kilcher did an analysis based on an open source language model called GPT-J-6B found that the bias was not as bad as he anticipated. However, It is also of paramount importance that the biases don’t create unnecessary skews in the AI models where such biases are not critical for the predictions, while acknowledging the fact that creating data sets without biases is a daunting task. The article goes on to discuss about how the biases in AI data collection may be controlled without violating the privacy of the sources.

 Complete article: New Scientist, June 25, 2022, page 16

Electron’s magnetic moment predicted from standard models and observations were improved with the new measurement made which is 2.2 times accurate than before. The new measurement is made at a precision of 1.3 parts per 10 trillion. The study was done at Northwestern University in Illinois. 

Complete article: New Scientist, October 22, 2022, page 09.

Sweeteners are widely used to replace sugars for low calory diet. Three such artificial sweeteners used in diet drinks, yoghurts and desserts are observed to dramatically halt the growth of multidrug-resistant bacteria, from a study at Brunel University in UK. The three sweeteners include, saccharin, cyclamate and acesulfame-K, successfully inhibited the growth of two bacteria Acinetobacter baumannii and Pseudomonas aeruginosa which are rampant in ICU ventilators and other medical environments. Resistance of bacteria to antibiotics is a natural process where bacteria evolves to resist the drug that challenges its very survival. However, the process is accelerated by overprescribing drugs in humans and misuse in animals, specially in hospital waste and hospital environment.

Complete article: Science Focus (BBC), January, 2023.

Hear diseases is one of the major contributor to deaths among humans worldwide and in Sri Lanka. Patients are  often administered statins that lower the cholesterol, but since heart diseases are not diagnosed early, those who could benefit from such medications do not receive them. Researchers based in Massachusetts General Hospital have developed a deep learning AI model to analyse the 10-year risk of a person of death due to heart attack or stroke from a single image of X-ray. The AI model is named CXR-CVD Risk included X-rays from 150,000 patients in a prostrate, lung, colorectal and ovarian cancer screening trials carried out by the National Cancer Institute. 

Complete article: Science Focus (BBC), January, 2023.

Guidelines of UK says, 14 units of alcohol a week, spread over a period of 3 days or more may not cause significant health issues. Although, how much people drink depends more on the price, availability and marketing, and not a health risk. 

Complete article: Science Focus (BBC), January, 2023.

Rivers in addition to transport and distribute water, the supply nutrients and sediments to downstream and oceans. While most anthropogenic activities lead to soil erosion (that elevate sediment and nutrient washoff in rivers), dam construction works on the contrary. This is further affected by the frequency and intensity of precipitation events, caused by climate change. A study by Deither et al. analysed sediment flux changes in 414 rivers worldwide using satellite data obtained between 1984—2020 (approximately 130,000 measurements).

Field measurements to study the dynamics of sedimentation characterization as the frequency of sampling would be too expensive, where image analysis could be beneficial, and irrespective of inaccessibility due to remoteness or terrain difficulty. The study found a 50% reduction in sediment flux in river of Northern Hemisphere, mainly as a result of entrapment of sediments in the dam. Simultaneously a 40% increase in suspended sediments was observed in the Southern Hemisphere, specially due to deforestation and mining and agricultural activities. Another modeling study on sediment retention in Mekong River showed that as little as 4% of the total sediment load is expected to reach the river delta. While dams and other measures of entrapment reduce sediment transport, agricultural and mining activities release much more than what has been trapped.

Changes in sediment fluxes have huge impact on the stability of river systems, their geomorphology, ecology and socioeconomic activities. Increased sedimentation may lead to ecosystem change, increase in flood risk, and increasing frequent dredging for removal of sediments and avail paths for navigation. In the meantime, decrease in sedimentation may starve ecosystems, affecting the food chain and socioeconomic activities that depend on them and aggravating erosion of sediments.

Environmental activities are always dynamic, and will change. However, sudden and frequent changes may have a dire effect on the entities discussed above. A clear strategy to manage sediment transport in rivers and subsequent ocean beds is therefore critical. For devising of effective management plans, studies are required to analyse the history, develop models to make predictions and the management protocols should be aligned to the predictions made.

Complete article: Science, June 24, 2022, page 1386.

Atop Indian Himalayas there lies a 4-metre liquid mirror telescope with a retractable roof(left) sits next to two optical telescopes. It appears as if someone scooped up a piece of the Bolivian salt flats, the world’s largest natural mirror, and put it in the Himalayas. The basin is part of a unique telescope. Situated in an observatory in the northern Indian state of Uttarakhand, the International Liquid Mirror Telescope (ILMT)uses the pool of shiny metal to gather light from the skies.

After a decade of works, ILMT, largest of its kind, and the first built to carry out astronomical observations was opened this year. According to Paul Hickson at the University of British Columbia in Canada, he states that “The idea was, if we can find a reflective liquid and rotate it with a very precise control system, it would focus light from above onto a detector and we could use that as a telescope”.

In the early 1900s, Robert Wood made a notable attempt but due to inability in maintaining the dish that held the metal spin at constant speed it failed. In1982, Ermanno Borra at Laval University in Quebec, Canada suggested dampening vibrations by pumping a thin layer of air between the basin that holds the mercury and the motor that rotates it, he also mentioned to pour the liquid resin on the dish surface first, letting it dry into the right shape, then pouring reflective liquid onto it as a coating, cutting down the amount of mercury needed. According to the suggestions of Borra and his colleagues, an experimental 2.7-metre-wide liquid mirror telescope near Vancouver, Canada was built. It was used to study the sodium layer in Earth’s atmosphere, but it wasn’t suitable for space observations, because of the cloudy weather near Vancouver.

Conventional instruments use glass mirrors to reflect light coming from stars and other cosmic objects. The mirror surface be polished smooth to a fraction of the wavelength of light which is almost impossible in large telescopes. One work around is making several smaller mirrors and then aligning them to form a big surface. The 4-metre ILMT cost around$2million, whereas the 3.6-metre optical telescope costs $17million. However, cost savings are irrelevant if we can’t build a working liquid mirror telescope.

There were several challenges in the project, getting enough mercury, coronavirus pandemic and so on. However, a strong determination kept the project alive. Unlike regular telescopes, which can point to different places, a liquid mirror telescope can only look straight up at the patch of sky overhead, which astronomers call the zenith. It was a good way of detecting objects that appear fleetingly, such as supernovae and passing asteroids. Another advantage of the ILMT is its location, right next to India’s biggest optical telescope. If astronomers spot something interesting through the ILMT, they can follow up with the other instrument.

Although these are cheaper to build liquid mirrors are tougher in construction. In the ILMT, the mirror dish floats on an air layer about10micrometres thick, much thinner than a human hair. Getting the resulting 4-metre-widedish up to the observatory, which is at elevation 2450 meters, was also quite an expedition. Finally, in early 2022, Hickson and Surdej achieved the first viable observation in the second week of May. The telescope is only operational for about eight to nine months a year, taking a break between July and October during monsoon season.

Observations so far include a rich stellar field in the Milky Way and NGC4274, a galaxy in the Coma Berenices constellation. The ultimate dream, for astrophysicist Anna Schauer at the University of Texas at Austin at least, is to build a giant liquid telescope on the lunar surface, to peer at the first stars in the universe. To do so, scientists would need a telescope with an incredibly large mirror. This would be practically impossible to manufacture and extremely expensive to put on the moon–unless it is made of ionic liquid. Such a huge project is likely to take decades to materialize. But the ILMT’s opening, after years of preparation, makes Schauer hopeful.

Complete article: New Scientist, December 10, 2022, page 41.

The article appeared on Scientific American discusses the ability of AI to write novel content in academic language with in-cite references in the right places and in relation to the right context. The author claims to have experimented with the AI called GPT-3, and discusses the issues on legalities and ethics of academic publishing and philosophical arguments on nonhuman authorship. THE GPT-3 is claimed to have writing creative pieces of article for example, opnion piece, a book on poetry and a new content from an 18^th century author. Although academics papers have been written on the ability of GPT-3, the author claimed to have observed none that had authorship of GPT-3.

The GPT-3 develops several arguments on a context and picks the best suitable, and the author has given prompts regarding the structure of the academic paper with no more human intervention.  The experiment was the AI to write an academic paper on AI itself, which was produced in less than two hours according to the article. The author discusses the questions he had when submitting the paper for publication regarding authorship and if authors had conflict of interest, and how he solved them. AI not not being sentient, asking if it had any conflict of interest raised concerns, the author claims. Several questions were raised in the article, for example;

1. Will journal editors require proof of human authorship in all publications? And that they have not used GPT-3 or similar AI.

2. If in case they used, does it need to be given authorship?

3. Could AI be ever given first author status?

It is time, future authorship of academic publication need to be revised.

Complete article: Scientific American, September, 2022, page 73.

Some Mercedes-Benz S-Class cars can drive themselves unaided, as long as you are in a specific car park in Germany. It is somewhat impressive, but disappointing given new Scientist proclaimed 15 years ago that a fully driverless car “may not be far off” The US start-up Cruise, which started providing a taxi service in San Francisco this year, is conducting what is arguably the most ambitious autonomous trial to yet. There are no drivers present, and the only people inside are customers. Everything has gone rather well, except for the sporadic stopping and blocking of traffic.

City streets can be challenging to navigate, so Cruise initially only offered rides at night, when streets were quieter and limited operation to San Francisco, which it has extensively mapped. Cruise initially only offered rides at night, when streets were quieter. It is, nonetheless, proof that driverless cars are possible and is a vital step to convincing people that the technology’s time is near. According to varying levels of car automation, as defined by the standards body SAE International level 0, levels 1,2, 3, and level 4 and level 5 indicates conventional car, driver-aided cars, full automation in “limited conditions” and total autonomy, all the time respectively.

Taxi services can sidestep problems by pausing operations if problems are spotted, without angering taxi drivers. Breckon says creating AI Making different models for every market is prohibitively expensive, so the software needs to be capable of handling all of it. Private car automation isn’t being entirely neglected, however, as is shown by Mercedes-Benz’s announcement this month that its S-Class model will now offer level 4automation, albeit only in one car park at Stuttgart airport, where it can park itself with help from sensors dotted around the area. The Tesla company also faces legal challenges over using the word “self-driving” in its marketing material. A level 4 car could still be more than five years from going on sale, and firms may never actually sell them to the public, says Breckon. If there’s a sudden spate of pedestrian deaths, they could have their autopilot function switched off, he says.

Complete article: New Scientist, December 17, 2022, page 13

Ocean covers more than 70% of the earth’s surface and as such, most rain fall over the oceans compared to land. However, lightening at sea is surprisingly rare which is revealed in a study as the salt spray getting in between clouds. The upper parts of the cloud freeze into a mixture of granular, rounded snow pellets called graupel and microscopic ice crystals. These ice and snow upon colliding with each other, transfer electric charges where the graupels tend to accumulate negative charges while the smaller ice crystals end up with positive charge.

Due to the updrafts of air, smaller ice crystals reach the top of the cloud while the graupels sink to the bottom, generating an electrical field between the top and bottom edges of the cloud. Depending on the charge difference, plausibility of a lightning event is triggered. When large water-absorbent particles of sea salt (from ocean spray) are present, the tiny droplets that condense on microscopic dust and soot to form clouds grow much more rapidly, becomes heavy enough to fall as rain well before the cloud forms tall enough. Although this idea was suggested before, a study by researchers from China, Israel and the U.S. experimentally found that the areas with salt spray had 90% less lightning. However, other factors also may contribute to the phenomenon such as atmospheric differences caused by local weather conditions such as wind and temperature, the researchers claimed.

Complete article: Scientific American, December, 2022, page 16.

More than 98% of Iran’s 1.648 million km² of land is subsiding. A subsidence at a rate of 4 cm/year is considered a crisis, while it was observed to be 6 cm/year in Iran, attributed to 25 years of water level decline. Major contributors were identified to be dam construction, climate change, inefficient water consumption by agriculture and industries and over extraction of groundwater. Subsidence will severely affect buildings, bridges, transportation lines and energy transmission in urban areas and can turn arable land in to dessert. Implementation of a integrated water resources management is critical in revival mission. 

Complete article: Science, June 17, 2022, page 6599.

Before we get to the discussion on the topic, let us talk a bit about cilium. Cilium is a membrane-bound organelle found on most types of eukaryotic cell and they do no exist in bacteria and archaea. It is of the shape of a tube protruding from the larger cell body. One piece of cilium may not be of much benefit, however, when in groups they contribute to  fluid flow. Cilia remove inhaled pathogens from respiratory tract, carry cerebrospinal fluid across brain cavities, transport eggs from the ovary to the uterus, and rain mucus from the middle ear to the nasal cavity.

Laboratory based artificial cilia are used in chips are expected to control fluids’ miniscule flow patterns. This concept currently in research status is believed by researchers to be used in testing blood, urine and other fluids to diagnose illnesses with efficiency more than that observed from the laboratory tests. In addition, this may only require much smaller volume of fluid samples compared to that required for laboratory testing. Studies from the past have attempted to make artificial cilia that can be moved using pressure, lightl, electricity and even magnets, while the major obstacle then was to produce tiny actuators, that triggered the motion. A study created cilium 1/20 of a millimetre long and 10 nm thick, with a strip of platinum on one side and a coating of titanium on the other. The article discusses the concept behind the functional attributes of artificial cilium, triggered by fluid flow. Running a low positive voltage triggers chemical reaction and frees oxygen from water molecules in the fluid. This results in bending of cilium, and is returned to original status once the voltage is reversed. Alternating this induces an oscillation in cilia, creating waves that move the fluid.

Complete article: Scientific American, September, 2022, page 18.

Turbulent swirls of water, ranging in size from a few kilometers to hundreds of kilometers across, peel off large ocean currents and mix heat and carbon dioxide into deeper ocean layers. They are the most energetic feature of the ocean, critical to getting climate models right yet largely invisible to satellites, except when they happen to sweep up a massive bloom of green phytoplankton.

After the launch of the Surface Water and Ocean Topography (SWOT) satellite, a joint venture between NASA and CNES, the French space agency, the ebb and flow of rivers and lakes will snap into focus. The satellite will capture eddies as small as 7 kilometers across and cover nearly the entire globe every 21 days. On land, SWOT will be able to map the changing height of more than 6 million lakes, from the Great Lakes down to ponds, capturing flows in rivers wider than 100 meters.

For nearly 4 decades, NASA and CNES have launched a series of radar altimeter satellites, which have monitored the accelerating rise of global sea levels, a basic indicator of climate change. With the aid of two 5-meter booms, each bearing an antenna to catch reflections of the radar signal SWOT pulses to Earth’s surface and gains a sharper view. With SWOT, seasonal changes in lakes and rivers can be precisely observed. SWOT will be able to chart how the levels of the world’s major rivers drop each time a dam or weir interrupts them, and how severely that fragments aquatic habitats. Moreover, SWOT will capture flood waters as they move downriver, which should help flood modelers.

Sönke Dangendorf, a physical oceanographer at Tulane University states that at the coastlines, SWOT will provide a detailed picture of how hot spots of sea level rise in the open ocean influence coastal inundation. Small eddies might warm nearby waters, fueling stronger hurricanes, thus trapping more heat at the depth too. This must be considered as there are many dwellers along the coasts

Complete article: New Scientist, December 17, 2022, page 13

Engineers in Germany have glued 18 solar panels to the outside of a huge tank of hot water for a German utility. It’s not the typical home for solar panels – most are flat, rigid silicon and glass rectangles arrayed on rooftops or at solar parks. Those are ultrathin organic films made by Heliatek, a German solar company. The goal is to expand solar power’s reach beyond flat land. “There is a huge market where classical photovoltaics do not work,” says Jan Birnstock, Heliatek’s chief technical officer. Organic photovoltaics (OPVs) such as Heliatek’s are more than 10 times lighter than silicon panels and cost just half as much to produce. Some are even transparent, which has architects envisioning solar panels, not just on rooftops, but incorporated into building facades, windows, and even indoor spaces. In the past few decades, researchers have come up with materials and designs that have reached efficiencies of nearly 20%, approaching silicon and alternative inorganic thin-film solar cells. OPVs allow them to tweak bonds, rearrange atoms, and mix elements from across the periodic table. Those changes represent flexibility to improve their materials’ ability to absorb sunlight, conduct charges, and resist degradation.

Solar energy is being added to the power grid at a massive rate, with more than 200 gigawatts coming online annually, enough to power 150 million homes. But solar and other green energy sources aren’t growing nearly fast enough to meet growing demand and forestall catastrophic climate change. Starting from global economic development, population growth, and the expected shift of much of the world’s cars and trucks from petroleum to electricity, the world’s electricity demand is expected to double by 2050. Countries must install renewables at four times the current pace to achieve global net zero carbon emissions by 2050.

All solar cells are sandwich-like devices, with semiconductors in the middle that absorb photons and convert them to electrical charges, which then migrate to metallic electrodes layered. When sunlight strikes silicon cells, the added energy kicks electrons out of their orbits around individual silicon atoms, freeing them. Each excited electron leaves behind an electron vacancy which flows to a cathode and electrons flow to the anode, creating an electric current. However, organic semiconductors tend to hold onto their charges more tightly. When OPVs absorb sunlight, there’s enough energy to kick an electron out of its atomic orbit, but not enough for the positive and negative charges to split up and move, these opposite charges stick to each other, creating what is known as an exciton. To generate electricity, the excitons must be separated into positive and negative charges that can travel to their respective electrodes. Over the decades, OPV researchers have sought to improve the performance of their devices by coming up with improved donors and acceptors. By incorporating soccer-ball-shaped carbon compounds called fullerenes into the materials efficiency above 5% was reachable. The fullerenes’ hunger for electrons makes them powerful acceptors. For the next decade, the action shifted to the donors. By 2012, a series of novel semi-conducting polymers used as donors propelled efficiencies to 12%.

A rival thin-film solar technology called perovskites came to progress. Perovskites are blends of organic and inorganic compounds that are cheap to make, easy to process, and great at capturing sunlight and converting it to electricity. While OPV progress stalled, the efficiency of perovskites skyrocketed from about 6.5% in 2012 to about 24% in 2020. Today, perovskites remain hot. But challenges with long-term stability and their reliance on toxic elements have sapped some enthusiasm. Meanwhile, OPVs soon got a burst of innovation of their own.

In 2015, researchers led by Xiaowei  Zhan reported the first of a new class of non-fullerene acceptors (NFAs) , it was only about 7% efficient. By 2016, new NFAs pushed OPV efficiency to 11.5%. By 2018, they hit 16%. And the records keep coming. Last year, Larson and his colleagues combined multiple donors, an NFA, and fullerene in a single layer, they created a material that enabled excitons to live longer which pushed its efficiency up to 18.4%. Silicon solar cells command an $85-billion-a-year market, with a 30-year track record and proven durability whereas OPVs remain niche products. U.S. startups Ubiquitous Energy and NextEnergy are developing energy generating OPV windows that primarily capture infrared photons while allowing visible light to pass through, something CIGS and other opaque thin films can’t do. r OPVs to become a significant source of green energy efficiency and durability must be concerned. Organics such as polymers and NFAs can be dissolved in solvents and machine-coated over large areas. But each layer in the sandwich like device must be completely smooth, with few or no imperfections, which can trap moving charges and reduce the overall efficiency. Controlling the makeup of the central layer of the sandwich containing the donors and acceptors. The intertwining ribbons of donors and acceptors must be extremely thin, because excitons created when photons strike the material can only migrate about 20 nanometers before the charges recombine and the opportunity to generate electricity is lost. Min’s team tailored a popular approach for manufacturing thin films at high speed called blade coating. Min’s team also calculates that the faster manufacturing rate could drop OPV costs more than 10-fold and make the price per kilowatt-hour (kWh) competitive with silicon. The major consideration should be whether such cells will retain the internal structure needed for high efficiency over decades. NFAs are especially susceptible, because the best ones consist of small molecules that can easily shift through the material. NFAs were replaced with acceptors woven into long polymers to help keep them in place. Hong Kong, reported all-polymer solar cells that had an efficiency of 17% and retained 90% of their efficiency under accelerated aging tests. However, under intense exposure to the ultraviolet (UV) in sunlight, the organics in solar cells can degrade. Forrest and his colleagues reported adding a thin layer of UV-absorbing zinc oxide—the same material in some sunscreens—to their OPV, which extended its life up to 30 years in accelerated aging tests.

OPVs already have a clear advantage over just about every other energy generating technology with a strikingly low carbon footprint to lighten further as their efficiency continues to set new records, lifetimes climb, and production methods advance. Those trends are buoying hopes of a world where solar power spreads along the curved facades of skyscrapers, the windows of the world which could make prospects for addressing climate CREDITS: change just a little bit brighter. 

Complete article: Science, November 11, 2022, page 588.