Sorry if this is spam... I'm drafting a paper about using a chatbot to control a molecular visualization viewer (I know this isn't new, it's just my contribution). I've written an evaluation benchmark with about 30 'tasks' that people may want to do with a structure viewer. The benchmark is evaluated using a few different tiers, with the final tier being a Vision Language Model (VLM) (e.g. LLM as a judge). However, I know I'm in a bubble, exploring only a small portion of the space of what people might want to do. I'd like to request volunteers to 1) input ideas to make the whole project more rounded, 2) help do human judging of the results (the VLM is pretty bad), 3) help draft new tasks to feed into the evaluation. I think the project as it stands makes a 'reasonable' CS conference paper, but to take it to the next level of publication, a wider expert validation is required. Anyone who's interested can get in touch directly or on the list. I've had good experience offering co-authorship for 'community contribution' to paper writing in the past, so if you'd like to help out, I have no problem adding you as a co-author on the paper. Cheers, Dan.

Where should I look for community these days? On Tue, 9 Jun 2026 at 08:27, Dan Bolser <dan.bolser@gmail.com> wrote: > Sorry if this is spam... > > I'm drafting a paper about using a chatbot to control a molecular > visualization viewer (I know this isn't new, it's just my contribution). > > I've written an evaluation benchmark with about 30 'tasks' that people may > want to do with a structure viewer. > > The benchmark is evaluated using a few different tiers, with the final > tier being a Vision Language Model (VLM) (e.g. LLM as a judge). > > However, I know I'm in a bubble, exploring only a small portion of the > space of what people might want to do. > > I'd like to request volunteers to 1) input ideas to make the whole project > more rounded, 2) help do human judging of the results (the VLM is pretty > bad), 3) help draft new tasks to feed into the evaluation. > > I think the project as it stands makes a 'reasonable' CS conference paper, > but to take it to the next level of publication, a wider expert validation > is required. > > Anyone who's interested can get in touch directly or on the list. > > I've had good experience offering co-authorship for 'community > contribution' to paper writing in the past, so if you'd like to help out, I > have no problem adding you as a co-author on the paper. > > > Cheers, > Dan. >

On Sat, Jun 13, 2026, 3:43 PM Dan Bolser <dan.bolser@gmail.com> wrote: > Where should I look for community these days? > > On Tue, 9 Jun 2026 at 08:27, Dan Bolser <dan.bolser@gmail.com> wrote: > >> Sorry if this is spam... >> >> I'm drafting a paper about using a chatbot to control a molecular >> visualization viewer (I know this isn't new, it's just my contribution). >> >> I've written an evaluation benchmark with about 30 'tasks' that people >> may want to do with a structure viewer. >> >> The benchmark is evaluated using a few different tiers, with the final >> tier being a Vision Language Model (VLM) (e.g. LLM as a judge). >> >> However, I know I'm in a bubble, exploring only a small portion of the >> space of what people might want to do. >> >> I'd like to request volunteers to 1) input ideas to make the whole >> project more rounded, 2) help do human judging of the results (the VLM is >> pretty bad), 3) help draft new tasks to feed into the evaluation. >> >> I think the project as it stands makes a 'reasonable' CS conference >> paper, but to take it to the next level of publication, a wider expert >> validation is required. >> >> Anyone who's interested can get in touch directly or on the list. >> >> I've had good experience offering co-authorship for 'community >> contribution' to paper writing in the past, so if you'd like to help out, I >> have no problem adding you as a co-author on the paper. >> >> >> Cheers, >> Dan. >> >