Cognir Ink.
Cognir Research
What it is: A research assist tool aimed at bridging the gap between curiosity/ideas and rigorous research direction.
Takes messy thoughts, ideas and concepts, something like:
"There's this idea that loneliness is as deadly as smoking. Is that real science or exaggeration? What does the evidence actually show for older adults?"
Or even something as messy as:
"I've been reading about how loneliness affects health outcomes in older adults, but like everything I find is either too narrow or too broad. There's that one study about inflammation markers but I can't remember who wrote it. Something about IL-6 and loneliness. And then there's the whole question of whether it's loneliness itself or the behaviors that come with it - like people who are lonely don't sleep well, don't eat well, maybe drink more. But then does that mean interventions should target behavior instead of loneliness? My advisor says I need a more focused question but I've been staring at this for three weeks and my brain is just static. Also someone mentioned something about how the effect size changes depending on how you measure loneliness - like UCLA"
Then, after determining the best and most feasible research direction from these ideas, prospective research questions are curated. A research question like the one below can be consistently expected from the tool:
"What is the dose-response relationship between weekly alcohol consumption and sleep disruption (measured by insomnia severity index) on cancer incidence in adults aged 40-65, and how does this compare to the dose-response relationship between smoking (measured by pack-years) and cancer incidence in the same population?"
Once a research question is selected, the tool makes use of a series of LLM calls for reasoning and web searches for grounding in real world research, curating a reading list for the researcher to get started with investigation.
The main advantage of using a tool like this for students and even for researchers would be that they can get oriented in new fields of research much more quickly. It promises to save the hours spent manually searching the web for literature, reading unrelated literature to orient oneself in a new field of research, and manually running searches through databases, not finding the best papers to get started with focused research.
For a research question like the one above, the tool would return a reading list as follows: