MSGID: 1:317/3 64927d25   
   PID: hpt/lnx 1.9.0-cur 2019-01-08   
   TID: hpt/lnx 1.9.0-cur 2019-01-08   
    AI reveals hidden traits about our planet's flora to help save species   
      
      
     Date:   
         June 20, 2023   
     Source:   
         University of New South Wales   
     Summary:   
         Machine learning can help extract important information from the   
         huge numbers of plant specimens stored in herbaria, say scientists.   
      
      
         Facebook Twitter Pinterest LinkedIN Email   
      
   ==========================================================================   
   FULL STORY   
   ==========================================================================   
   In a world-first, scientists from UNSW and Botanic Gardens of Sydney,   
   have trained AI to unlock data from millions of plant specimens kept in   
   herbaria around the world, to study and combat the impacts of climate   
   change on flora.   
      
   "Herbarium collections are amazing time capsules of plant specimens,"   
   says lead author on the study, Associate Professor Will Cornwell. "Each   
   year over 8000 specimens are added to the National Herbarium of New South   
   Wales alone, so it's not possible to go through things manually anymore."   
   Using a new machine learning algorithm to process over 3000 leaf samples,   
   the team discovered that contrary to frequently observed interspecies   
   patterns, leaf size doesn't increase in warmer climates within a single   
   species.   
      
   Published in the American Journal of Botany, this research not only   
   reveals that factors other than climate have a strong effect on leaf   
   size within a plant species, but demonstrates how AI can be used to   
   transform static specimen collections and to quickly and effectively   
   document climate change effects.   
      
   Herbarium collections move to the digital world Herbaria are scientific   
   libraries of plant specimens that have existed since at least the 16th   
   century.   
      
   "Historically, a valuable scientific effort was to go out, collect plants,   
   and then keep them in a herbarium. Every record has a time and a place   
   and a collector and a putative species ID," says A/Prof. Cornwell,   
   a researcher at the School of BEES and a member of UNSW Data Science Hub.   
      
   A couple of years ago, to help facilitate scientific collaboration,   
   there was a movement to transfer these collections online.   
      
   "The herbarium collections were locked in small boxes in particular   
   places, but the world is very digital now. So to get the information   
   about all of the incredible specimens to the scientists who are now   
   scattered across the world, there was an effort to scan the specimens to   
   produce high resolution digital copies of them."  The largest herbarium   
   imaging project was undertaken at the Botanic Gardens of Sydney when   
   over 1 million plant specimens at the National Herbarium of New South   
   Wales were transformed into high-resolution digital images.   
      
   "The digitisation project took over two years and shortly after   
   completion, one of the researchers -- Dr Jason Bragg -- contacted me from   
   the Botanic Gardens of Sydney. He wanted to see how we could incorporate   
   machine learning with some of these high-resolution digital images of the   
   Herbarium specimens."  "I was excited to work with A/Prof. Cornwell in   
   developing models to detect leaves in the plant images, and to then use   
   those big datasets to study relationships between leaf size and climate,"   
   says Dr Bragg.   
      
   "Computer vision" measures leaf sizes Together with Dr Bragg at the   
   Botanic Gardens of Sydney and UNSW Honours student Brendan Wilde,   
   A/Prof. Cornwell created an algorithm that could be automated to detect   
   and measure the size of leaves of scanned herbarium samples for two plant   
   genera -- Syzygium(generally known as lillipillies, brush cherries or   
   satinas) and Ficus(a genus of about 850 species of woody trees, shrubs   
   and vines).   
      
   "This is a type of AI is called a convolutional neural network, also   
   known as Computer Vision," says A/Prof. Cornwell. The process essentially   
   teaches the AI to see and identify the components of a plant in the same   
   way a human would.   
      
   "We had to build a training data set to teach the computer, this is a   
   leaf, this is a stem, this is a flower," says A/Prof. Cornwell. "So we   
   basically taught the computer to locate the leaves and then measure the   
   size of them.   
      
   "Measuring the size of leaves is not novel, because lots of people have   
   done this. But the speed with which these specimens can be processed and   
   their individual characteristics can be logged is a new development."   
   A break in frequently observed patterns A general rule of thumb in the   
   botanical world is that in wetter climates, like tropical rainforests, the   
   leaves of plants are bigger compared to drier climates, such as deserts.   
      
   "And that's a very consistent pattern that we see in leaves between   
   species all across the globe," says A/Prof. Cornwell. "The first test we   
   did was to see if we could reconstruct that relationship from the machine   
   learned data, which we could. But the second question was, because   
   we now have so much more data than we had before, do we see the same   
   thing within species?"  The machine learning algorithm was developed,   
   validated, and applied to analyse the relationship between leaf size   
   and climate within and among species for Syzygiumand Ficusplants.   
      
   The results from this test were surprising -- the team discovered that   
   while this pattern can be seen between different plant species, the   
   same correlation isn't seen within a single species across the globe,   
   likely because a different process, known as gene flow, is operating   
   within species. That process weakens plant adaptation on a local scale   
   and could be preventing the leaf size-climate relationship from developing   
   within species.   
      
   Using AI to predict future climate change responses The machine learning   
   approach used here to detect and measure leaves, though not pixel perfect,   
   provided levels of accuracy suitable for examining links between leaf   
   traits and climate.   
      
   "But because the world is changing quite fast, and there is so much   
   data, these kinds of machine learning methods can be used to effectively   
   document climate change effects," says A/Prof. Cornwell.   
      
   What's more, the machine learning algorithms can be trained to identify   
   trends that might not be immediately obvious to human researchers. This   
   could lead to new insights into plant evolution and adaptations, as   
   well as predictions about how plants might respond to future effects of   
   climate change.   
      
       * RELATED_TOPICS   
             o Plants_&_Animals   
                   # Nature # Endangered_Plants # Botany   
             o Computers_&_Math   
                   # Computer_Modeling # Computers_and_Internet #   
                   Neural_Interfaces   
             o Fossils_&_Ruins   
                   # Early_Climate # Ancient_DNA # Evolution   
       * RELATED_TERMS   
             o Computational_neuroscience o Scientific_visualization o   
             Data_mining o Paleoclimatology o Computer o DNA_microarray o   
             Spinach o Chloroplast   
      
   ==========================================================================   
   Story Source: Materials provided by   
   University_of_New_South_Wales. Original written by Lilly Matson. Note:   
   Content may be edited for style and length.   
      
      
   ==========================================================================   
   Journal Reference:   
      1. Brendan C. Wilde, Jason G. Bragg, William Cornwell. Analyzing   
         trait‐climate relationships within and among taxa using   
         machine learning and herbarium specimens. American Journal of   
         Botany, 2023; 110 (5) DOI: 10.1002/ajb2.16167   
   ==========================================================================   
      
   Link to news story:   
   https://www.sciencedaily.com/releases/2023/06/230620113755.htm   
      
   --- up 1 year, 16 weeks, 1 day, 10 hours, 50 minutes   
    * Origin: -=> Castle Rock BBS <=- Now Husky HPT Powered! (1:317/3)   
   SEEN-BY: 15/0 106/201 114/705 123/120 153/7715 218/700 226/30 227/114   
   SEEN-BY: 229/110 112 113 307 317 400 426 428 470 664 700 291/111 292/854   
   SEEN-BY: 298/25 305/3 317/3 320/219 396/45 5075/35   
   PATH: 317/3 229/426