Anveshan – A Model for Search

Bharat Banavalikar, Anvita Bhat, Apoorva Joshi, Praveen Talavar, Prakash Hegade
2020 Procedia Computer Science  
Search and the engines performing them have witnessed significant transformations over time. While the web wants to be semantic, search demands to be customized, and every attached entity wants to learn; the challenges magnify leaving behind intricate woven questions to be addressed and resolved. The concept of machine learning challenges the designer to exploit the full potential of a machine capability. Through this paper, we propose a model -Anveshan, which uncovers the said challenges by
more » ... ucing a learning model. The model carries hooks to establish the parameters, uses the user-defined initial weights, and the humongous data available on the web. It then prepares a list of results based on the considered parameters. The results learn and rearrange based on user clicks and preferences. The model also produces a bit string for the considered hooks, indices it, and makes it agent readable. The intention to create a universal bit string is to make it machine-understandable and contribute towards the semantic web. The paper discusses the model and then presents a specific case study in results and discussion. Anveshan works towards creating a learnable exploration space as per the user specifications. Anveshan is a fork-able model overlaid with a custom environment. The results are promising to guide towards a better search space. Abstract Search and the engines performing them have witnessed significant transformations over time. While the web wants to be semantic, search demands to be customized, and every attached entity wants to learn; the challenges magnify leaving behind intricate woven questions to be addressed and resolved. The concept of machine learning challenges the designer to exploit the full potential of a machine capability. Through this paper, we propose a model -Anveshan, which uncovers the said challenges by inducing a learning model. The model carries hooks to establish the parameters, uses the user-defined initial weights, and the humongous data available on the web. It then prepares a list of results based on the considered parameters. The results learn and rearrange based on user clicks and preferences. The model also produces a bit string for the considered hooks, indices it, and makes it agent readable. The intention to create a universal bit string is to make it machine-understandable and contribute towards the semantic web. The paper discusses the model and then presents a specific case study in results and discussion. Anveshan works towards creating a learnable exploration space as per the user specifications. Anveshan is a fork-able model overlaid with a custom environment. The results are promising to guide towards a better search space.
doi:10.1016/j.procs.2020.04.256 fatcat:37egmo2b2nevhgqfpnfvo4ryz4