Course:CONS200/2021/ Artificial Intelligence (AI) and Citizen Science Working Together in Ecological and Wildlife Data Collection
Artificial Intelligence (AI), the study and development of computer systems that can copy intelligent human behavior, is a technology that the majority of people have heard of and might have a general understanding of. The science and term has gained popularity since its inception in 1956 at a Dartmouth College workshop. Citizen science, allows for people with interest to engage in data collection within their hobby’s field. Although not as recognizable of a term, it is thrown around regularly by media outlets, and people involved in the field of science. These two terms may seem dissimilar, one being human hobby based and another being a complex science often glorified by Hollywood, but these two developments are being partnered to improve Wildlife and Ecological data collection moving forward. It is important to understand AI and Citizen Science, as well as the more traditional approaches to data collection in order to see these ideas as assimilative.
Overview of Artificial Intelligence (AI)
Artificial Intelligence (AI) describes the development of computer systems that are actively learning or collecting data. Artificial Intelligence can be utilized in the study of Conservation to accurately retrieve data on local wildlife populations, ecosystem health, or climate monitoring. Artificial Intelligence is commonly used to make accurate predictions. An example of this is the autopilot Artificial Intelligence that has been implemented in Tesla vehicles to avoid crashing in the emergency that the driver cannot operate the vehicle. An instance of Artificial Intelligence being used in Conservation is a compilation and collection of bird calls into databases which monitor wildlife, and the number of bird species in a local area. By allowing Artificial Intelligence to become easily accessible to the public through an application on a smart phone. Citizens are then able to participate in data collection and can provide necessary data for experts. Artificial Intelligence can process data much more efficiently and effectively than humans. This is seen through comparing data sets and the transfer of data.
Applications of Artificial Intelligence
A challenge faced by Artificial Intelligence is learning, which is a common way to gain knowledge through experience. The problem with current learning algorithms is that the artificial intelligence believes that previous data will presumably continue. Another reason is because it takes time to develop an accurate machine learning mechanism that can interpret new data. This has been seen in social media where an artificial intelligence can track your data to recommend other search results or products similar to other ones you've used before. Statistical analysis is a strong suit for Artificial intelligence since, it is able to precisely and efficiently collect data. This can be proven to be beneficial when combined with Citizen Science to provide quality data for conservation of wildlife species. Artificial Intelligence can also be applied in Computer Science. Computer scientist study how programs, functions, coding, and learning algorithms can operate on computers. Perceptions of Artificial Intelligence are being implemented in mobile devices and everyday objects. Sensors such as face or touch identification in iPhones or inside tea kettles that notify you when the water has become hot. The perception of AI is beginning to pervade into the daily lives of individuals.
History of Artificial Intelligence
The term AI was coined by Allen Newell, Herbert Simon, John McCarthy, and Marvin Minsky in the 1950’s. Computer systems were used to solve complex math equations and prove theorems using simulations. These equations were solved in a simple true or false form. In games such as chess, the algorithm is able to learn from past games and utilize previously attained skills in new situations, until it has knowledge in every possible move set, making it virtually impossible to beat. In today's world, AI uses much more powerful computer systems for more complex machine learning and data collection. Artificial Intelligence can also be used in education, research, learning, and educating individuals on the possibilities of Artificial Intelligence. Some people believe that machine learning may accelerate exponentially, which AI researchers will need to monitor, and possibly apply a limit to how much it can learn. In today's society, active research is being put into refining the capabilities of Artificial Intelligence to make it fundamentally more intelligent.
Overview of Citizen Science (CS)
Citizen science allows for public engagement in data collection. It is the involvement of the public in scientific research. The exact definition is under contention, as differentiating volunteers, amateurs and citizen scientists can be more difficult than expected. Some classify by quantity and quality of public engagement, others define citizen science by who is leading the project, and how the project is set up. The precise definition is less important, as we want to focus on the process and effectiveness of citizen science.
There is a gap between the science community and the public . Citizen science has the capability bridge that gap and make it shorter. Scientists work at the for front of issues that face the globe and people are not, and do not know how to engage. Considering science is often publicly funded, the resilience of the notion that peer evaluation/ assessment is most important is outdated and outrageous. None the less this is where citizen science can actively provide an avenue between the world of scientist and researchers to ordinary people. For example, students, when they are introduced to different modes of citizen science or directly to the scientists conducting research, they gain a higher interest in projects. In addition, after working on projects citizen science participants will bring information back to their communities and relay information to community members, creating and growing interests in scientific topics. The easiest way to create public interest in the science community is to have people participate in it, people are inherently more interested in things they have invested time and effort into.
One worry often articulated by scientists working with citizens is on the quality of data. Some feel worried about the accuracy of the data being recorded, others worry if there is a big enough level of acceptance within the scientific community to see it as a legitimate data collection method.
Overcoming the Obstacles of Data Collected by Citizen Scientists
- Training / Supervision
- Cross Check Results
- Quiz / Test level of knowledge
- Simplify Tasks
In order for citizen science to be trusted the data collected must be true and legitimate. These strategies can, and are used to legitimize data collected under citizen science in order to create legitimacy and trust in the scientific community.
Ecological and Wildlife Data Collection
In research there are different methods used to gather information. The two commonly known categories are primary and secondary data. Primary data is collected for the first time by a researcher, while secondary data has already been collected or produced by others.
Primary Data Collection
For a researcher, primary data collection is costly, time consuming, and requires money. Primary data sources include surveys, observations, experiments, questionnaires or personal interviews. Primary data can be divided into two categories; Quantitative and Qualitative.
Commonly deals with numbers, quantities and values. This type of data is used for analytics because it is numeric and measurable. It is also an objective type of data and is seen as more reliable then qualitative data.
Methods for gathering quantitative data include, using logical or statistical observations to draw conclusions. Quantitative data aims at establishing cause and effect relationships between two variables by using mathematical, computational and statistical methods.
This form of observation deals with numbers and measurements, used for objects or phenomena that can be tracked and measured with precision. Large data samples help researchers find possibly significant trends and patterns within a general population. This is where citizen science has been proven useful.
Most often focuses on personal description rather then numeric values. It is less reliable then quantitative data and is difficult to measure. However, this data helps reveal the motivative behind quantitative data results.
Methods for gathering qualitative data includes, verbal narrative like spoken or written data. For example, case studies, grounded theory, ethnography, historical and phenomenology are types of qualitative research.
Secondary Data Collection
For a researcher, secondary data is more accessible, saves more time and money in comparison to primary data collection. Secondary data collection is found in many forms for example, it can be a journal, article or website. It can also be a consensus, government publication or internal records from an organization.
Wildlife Data Collection
Research by Sullivan et al. (2017, p. 12) indicates that data collected through citizen scientist initiatives can be used for conservation action. This new type of information acts as a compliment towards existing data gathering methods i.e. citizen science projects where people collect real-time observations on a species occurrence and abundance. These types of projects use the power of the internet and the crowd of observers to gather data at a large volume and scale. Citizen based observational data currently serves as the fastest way to understand species distribution. Through, the opportunities provided by monitoring a broad range of species over a long period of time.
Citizen Science Working with AI
The pairing of AI and citizen science allows for researchers, in the fields of ecology and wildlife, to have greater access to diverse streams of data from citizen scientists all over the world. These citizen scientists can provide surpluses of data for researchers, saving valuable time.
Protein Structure Prediction and Citizen Science
A game has been developed by scientists that allows the average person to contribute to significant scientific discoveries. The game is called Foldit, the goal is to advance the ability of protein structure prediction by having people solve puzzles which resemble real folded proteins. One of the goals of the project is to see if humans are better at folding proteins than computers, and then to teach the human derived pattern recognition to computers, thusly increasing the speed at which proteins can be analyzed.
Why is this important?
Knowing the structures of proteins in key to understanding how they work. Many malicious diseases such as HIV/AIDS, cancer, and Alzheimer's all have direct relations to proteins. Foldit allows the average citizen to contribute to the advancement of science, by solving protein based puzzles, thusly helping to advance the understanding of diseases and their fields of research.
Birdwatching Data and Artificial Intelligence
A particular field which shows promise in pairing AI with citizen science is the field of ornithology(study of birds). Numerous organizations which are involved in bird watching, provide applications to citizen scientist for data collection, with a symbiotic benefit of both parties. Bird watchers get the benefit of storing their birding lists and sighting locations, while researchers get access to new and constant data on bird sightings. These emerging technologies harness the average individuals' interests in nature to capture data. The Audubon Society, Cornell Bird Lab, and others have paired with a platform called eBird, which is an open access observational research and conservation application, where users are able to record bird sighting data. Bird sighting data, from citizen scientists, is reviewed, and then processed by Ai to compile migration maps and visualizations, which are then used for science and conservation.This platform uses primary and secondary data collection methods. But it mainly relies on observational and monitoring research data, collected from its users.
“…eBird…aims to collect data on all bird species, year-round (i.e., across breeding, migration, and non-breeding periods), from any location on the planet, and make these data openly available for research, education, and conservation. Currently, eBird serves as a major source for observational data on bird occurrence, providing roughly 20% of the data available in the Global Biodiversity Information Facility (GBIF)…an open access biodiversity data clearinghouse.”
eBird seeks to meet three major challenges:
There are limited mechanical computational systems in the field of organism classification. Humans are able to easily identify hundreds of species, which is an incredibly complex task if you consider identifying birds visually with a glimpse, hearing a birds song, then cross referencing that information with the time of year and geographic location.
Species identification can not yet be automated by computers, species identification must still be done by humans, causing eBird to rely on crowdsourcing techniques. In order to maximize participation, eBird uses tools to appeal to the birding community, which are also useful for scientists studying birds. These features include: keeping track of a users bird record, sorting personal bird lists by date and region, sharing bird lists with others, and visualizing a users observations on maps and graphs. By maximizing the rewards to participants, eBird ensures that participation readily occurs on their platform.
Identifying Synergies Between Humans and Machines
The first hurdle in using citizen data is the inevitability of users misidentifying birds. The second hurdle is the uneven distribution of eBird users geographically. The third hurdle is the individuals ability to identify birds and the wide range of expertise among participants. As there is a high probability of data quality issues, there must be a system in place to vet data. In order to automate this data review process eBird screens data based on the frequency a species is reported within a period of time, to determine the possibility that the species was correctly reported at any spatial level.
Bald Eagle Population Tracking
The bald eagle is a culturally significant species in North America which has seen considerable population declines, like many other bird species, since the industrialization of the continent. Due to the significance of the species, many conservation initiatives have been directed towards the rejuvenation of bald eagles over that last 50 years. Through the collaboration of citizen scientists using the previously mentioned platform, ebird, it was determined that count of bald eagles has quadrupled since the previous population report published in 2009. This places estimates of bald eagle populations to be approximately 316,708 eagles across the United States. The new USFWS report suggests nesting pairs of bald eagles have doubled since the 2009 report, 71,467 pairs, a significant increase from the all time low documented nesting eagle pairs, which was 417 known eagle nests in 1963. The recovery of the species, and many others, is tracked through the reported sightings of individuals. By having citizen scientists contribute their sighting data, which is then compiled by AI and accessible to governmental and non governmental organizations, affective decisions on further conservations steps can be made .
Implementing Nest Box Location Using eBird Data
The Rocky Mountain Naturalist club used data from eBird to discern the placement of new duck nest boxes at Elizabeth Lake in Cranbrook, BC. The lake is an eBird hotspot with 208 species observed and over 1500 completed checklists as of 2021. The club used eBird data to gather information on different bird species, migration and nesting phenology. This information educated the club about which duck species were present and the proper nesting boxes that should be used. Nesting boxes help increase the breeding success of many species of cavity-nesting waterfowl.
Citizen Science is the ability for ordinary people to take part in research, whilst Ai is a computational system developed to actively learn. The two seemingly different methods working together is not a technology for the future, it is very real and happening across scientific fields today. All over the word citizen scientists are becoming part of programs, such as ebird, which are making significant contributions to ecological data collection and shifting the paradigms of how data can be collected. Artificial intelligence is proving to be indispensable in the field of data analysis, as it is able to use data in powerful ways to make predictions. By tuning aspects of Artificial Intelligence such as learning, planning, and general intelligence, it will allow the AI to inevitably perceive data intelligently. Ai can also be used to vet data collected by citizen scientists which has a high probability of being false, due to the lack of professional expertise in the respective field, further empowering the science behind Ai, working together with citizen scientists to revolutionize the way data can be collected in the fields of ecology and wildlife management.
Please use the Wikipedia reference style. Provide a citation for every sentence, statement, thought, or bit of data not your own, giving the author, year, AND page. For dictionary references for English-language terms, I strongly recommend you use the Oxford English Dictionary. You can reference foreign-language sources but please also provide translations into English in the reference list.
- Oxford University, P (2021). "Artificial intelligence". Oxford Learner's Dictionary.
- Crevier, Daniel (1993). "Ai : the tumultuous history of the search for artificial intelligence / Daniel Crevier". Basic Books.
- Riesch, H (2014). "Citizen science as seen by scientists: Methodological, epistemological and ethical dimensions". SAGE Publications.
- Dick, Stephanie (July 1 2019). "Artificial Intelligence". doi:https://doi.org/10.1162/99608f92.92fe150c Check
|doi=value (help). Check date values in:
- "Autopilot". Tesla.
- Kwok, Roberta (March 4 2019). "AI empowers conservation biology". Check date values in:
- Zeng, Daniel; Chen, Hsinchun; Lusch, Robert; Li, Shu-Hsing (30 December 2010). "Social Media Analytics and Intelligence". IEEE Xplore. doi:10.1109/MIS.2010.151.
- Fang, Fei; Tambe, Milind; Dilkina, Bistra; Plumptre, Andrew, eds. (2019). Artificial Intelligence and Conservation. pp. 7–11. ISBN 978-1-316-51282-0 Check
|isbn=value: checksum (help).
- "The Power of Citizen Science". CITIZENSCIENCE.ORG.
- Poliakoff, Ellen (December 2007). "What Factors Predict Scientists' Intentions to Participate in Public Engagement of Science Activities?". Sage Publications.
- Doyle, Cathal (2019, April 7). "Using the Web for Science in the Classroom: Online Citizen Science Participation in Teaching and Learning". Teaching and Learning Research Initiative. Check date values in:
- S, Surbhi (July 13, 2020). "Difference Between Primary and Secondary Data". Key Differences. Retrieved March 28, 2021.
- S, Surbhi (November 19, 2018). "Difference Between Qualitative and Quantitative Research". Key Differences. Retrieved March 28, 2021.
- Nelson, D (March 4, 2020). "Quantitative Observation: Definition And Examples". Science Trends. Retrieved March 28, 2021.
- Sullivan, B. L., Phillips, T., Dayer, A. A., Wood, C. L., Farnsworth, A., Iliff, M. J., Davies, I. J., Wiggins, A., Fink, D., Hochachka, W. M., Rodewald, A. D., Rosenberg, K. V., Bonney, R., & Kelling, S. (2017). Using open access observational data for conservation action: A case study for birds. Biological Conservation, 208, 5–14. https://doi.org/10.1016/j.biocon.2016.04.031
- Kelling, S., Gerbracht, J., Fink, D., Lagoze, C., Wong, W.-K., Yu, J., Damoulas, T., & Gomes, C. (Spring 2013). "A human/computer learning network to Improve Biodiversity conservation and research". AI Magazine. 34(1). doi:https://doi.org/10.1609/aimag.v34i1.2431 Check
|doi=value (help).CS1 maint: multiple names: authors list (link)
- Mortier, Saira (2020). "The Science Behind Foldit". Foldit Solve Puzzles for Science. Retrieved 14-04-2021. Check date values in:
- "About eBird". eBird. Retrieved February 24, 2021.
- eBird. "About". eBird. Retrieved April 15, 2020.
- Axelson, Gustave (March 24th 2021). "Bald eagle count quadruples, thanks in part to eBird data boost". Cornell Chronicle. Retrieved April 5th 2021. Check date values in:
- eBird (April 16, 2021). "Elizabeth Lake, Cranbrook". eBird. Retrieved April 16, 2021.
- En.wikipedia.org. (2018). Writing better articles. [online] Available at: https://en.wikipedia.org/wiki/Wikipedia:Writing_better_articles [Accessed 18 Jan. 2018].
|This conservation resource was created by Course:CONS200. It is shared under a CC-BY 4.0 International License.|