Context
For 2+ years, I was into hardcore academic research. Currently, I'm working in corporate, but a large amount of time is still spent on going through research works (papers, articles or sometimes even book chapters) in a regular basis. Google Scholar is good, but it doesn't help in reducing the time spent in finding relevant papers and to understand whether that paper can be of any use or not. There are some tools available online, but they were not very useful to me.
Probable Solution?
I am of the opinion that good visuals can help you understand complex patterns easily and fast. I have done some R&D on this, and with the current advancements in AI models, I think a good solution can be made.
Currently, I've built a prototype in my jupyter notebook. Basically, a user can provide any text query (like you do in Google Scholar search bar), based on that query, I scrap a few data sources, use a few APIs to get a good list of research papers that might be relevant. This is what Google Scholar also does.
But, after that I build a Knowledge Graph with all those papers. I try to find the relations between those papers. Are there common authors? (which may translate to important researchers in a topic), Are there any common themes/ key concepts in those listed papers? What is the chronology of the relevant research papers? etc.
Conclusion
I built this in my local environment and changing/ tweaking this according to my needs or questions that I want answered out of this. But I want to understand,
For 2+ years, I was into hardcore academic research. Currently, I'm working in corporate, but a large amount of time is still spent on going through research works (papers, articles or sometimes even book chapters) in a regular basis. Google Scholar is good, but it doesn't help in reducing the time spent in finding relevant papers and to understand whether that paper can be of any use or not. There are some tools available online, but they were not very useful to me.
Probable Solution?
I am of the opinion that good visuals can help you understand complex patterns easily and fast. I have done some R&D on this, and with the current advancements in AI models, I think a good solution can be made.
Currently, I've built a prototype in my jupyter notebook. Basically, a user can provide any text query (like you do in Google Scholar search bar), based on that query, I scrap a few data sources, use a few APIs to get a good list of research papers that might be relevant. This is what Google Scholar also does.
But, after that I build a Knowledge Graph with all those papers. I try to find the relations between those papers. Are there common authors? (which may translate to important researchers in a topic), Are there any common themes/ key concepts in those listed papers? What is the chronology of the relevant research papers? etc.
Conclusion
I built this in my local environment and changing/ tweaking this according to my needs or questions that I want answered out of this. But I want to understand,
- Whether this solution has any value to anyone else?
- What else can I add to make this more useful?