Undermind

See how to search for complex research topics with Undermind:

?????????????????? ?????? ???????????? ???????????????????? ???????????? ???? ?????????? ?????? ????????????????????? How much time do we spend on the literature search? Actual systems, including Perplexity and ConsensusAI, do not find information regarding proper scientific literature. Former Massachusetts Institute of Technology scientists developed Undermind! The results are insane. You get the most cited article and a report about the match. Results (publications) can be exported as *RIS, *CSV, *BIBL. For now, it searches through the abstracts and metadata of more than 200M articles, but it will also look for the text within the articles. 5 Searches are free.. If you find the helpful post, please repost and follow for more.

Finding relevant papers for literature review takes a lot of time. So, MIT researchers built Undermind, an AI-powered search engine. It can generate well-researched overviews on any topic and is 10-50 times better than Google Scholar. Here's how to use it: 1. Go to undermind[.]ai and sign up for a free account. 2. Once you've logged in, this is how Undermind would look like. Type in the topic of your research project. You can be as vague as possible. Undermind will you a brief overview about your topic. Then it will ask you to narrow down your topic. 3. Once you've narrowed down your topic, type in your question, and press Enter. Undermind will reword your question. Then scroll down a little and click on "Submit Deep Search." Undermind will take 3-5 minutes to compile a comprehensive answer to your question. 4. Once it's done, Undermind will give you a detailed overview of the project with citations to published papers. Hover your cursor over a citation and it will show you the relevent article. 5. Undermind will also show you different categories of papers relevant to your topic. This can help you a lot in framing your project. 6. Undermind also has an option for timeline of published papers and the citation network. I didn't find the user interface of the timeline very helpful. But I am sure this will improve in the future. 7. If you scroll all the way down, Undermind will give you a list of references. You can enable abstracts and it will show you abstracts of all the papers in the list. It will also give you a score of how relevant a given paper is to your topic. 8. You can also ask it to write brief review articles on a given topic. After you have run a search, ask it to give you a review article. These brief articles can be a great way to familiarize yourself with new topics. 9. Undermind will save all your searches. To look at your previous searches, simply click on "History." Found this post on Undermind helpful? 1. Repost to share it with your friends and colleagues. 2. Follow me for more posts on how to use AI apps for academic purposes.

I thought this would be common knowledge by now, but many of my scientist and researcher friends were pleasantly surprised when I shared this: ?????? ???????????????????????? ?????? ?????????????????????????????? ?????? ?????? ???? ???????????????? ???????????????? Traditionally, graduate students (myself included at one point) relied heavily on subject-matter experts, such as advisors, who could recommend papers from their vast memories. From there, it often took days or weeks to sift through articles to pinpoint relevant citations and build a coherent narrative. But now, tools like Undermind, an LLM-based prompt engineering startup created by MIT graduates, are reshaping this process. Here’s how a modern, closed-loop research workflow might look: 1. Ask a research question. Start by framing an inquiry or literature review topic with Undermind. Refine the question through dialogue with Undermind if necessary. 2. Source journal articles. Let Undermind pull relevant articles from pre-print databases like arXiv and many partnering publishers, complete with summaries tailored to the research question. 3. Dive deeper with OpenAI's GPT or Anthropic's Claude. Upload the most relevant articles to these LLMs and ask the refined research question to generate a synthesized literature review. 4. The result? ?? ???????????????????? ???????????? ???? ?????????? ?????????????? ???? ??????????. 5. Of course, don't forget to check the results against common sense, first principles, and/or feedback from a subject-matter expert!

Proud to be on the editorial board for Resource Reviews for Katina Librarianship Elevated (successor to the 25 year old Charleston Advisor) published by Annual Reviews and officially launched at the Charleston Conference today. With the influx of new AI products and features entering the academic market, our resource reviews have naturally evolved to cover them. At the time of writing, we have product reviews of Statista Research AI, Undermind, and SciSpace as well as other product reviews of interest to the academic market. https://lnkd.in/g6M3KS_k

SMU libraries has subscribed to Undermind.ai. Just register with an account registered with SMU email and you will automatically be upgraded to premium version. Unlike other search (even enhanced with "AI", transformer based LLM) tools Undermind (https://undermind.ai/home/) is a agent-like tool that mimics human researcher behavior by running a initial search, and based on the results adapting for other searchs or doing citation searching, ultimately resulting in some of the most relevant results I have seen for an academic search tool (including those using "AI") so far. As an aside the opensource PaperQA2 project has a similar concept at least on the retrieval side. Then again you do have to wait 3 minutes as opposed to a few second for most tools, so the comparison isn't quite fair. Feedback from our community has been extremely enthusiastic. A Phd refers to using Undermind like transitioning from using "log tables to a electronic calculator". A faculty who tested it in a well researched area noted it found roughly 89% of papers, though I think this is highly subject dependent (It probably does better in CS and STEM). Our own testing against gold standard sets from systematic reviews in more social sciences areas show it finding only 30-40% of gold standard set (partly because it doesn't have full text and evaluates on title abstracts only) but it's still finds 10-20% more than other AI tools we tested. I have noticed and seen reports that it sometimes totally falls down badly on business/management type topics, sometimes due to source reasons (HBR type content is restrictive), other times for unknown reasons. I personally use it now for trying to help confirm if a certain specific type of research DOESN'T exist, if Undermind can't find anything relevant it might still exist but the chances against it goes way up. It can also help you quickly scope out how much if any work has been done in the area you are wondering about. https://undermind.ai/home/ edit : Added more of the short comings

The best use case of #AI in research (thus far) is for conducting #literaturereviews. AI enables researchers to find more relevant papers in less time without any risk of hallucinations. I'm delighted to invite #librarians and #researchers to a **free** workshop on my 3 favorite AI tools for literature review on October 14 at 1:00 PM EST. I will show participants how to get the most out of these tools and then there will be time for guided practice + feedback. A recording will be sent to those who can't attend live. To register: https://lnkd.in/eRkDz3G4 Academic Language Experts Dimensions, Digital Science, Sarah Greaves, Julia Hawks, Ben McLeish, Undermind, Joshua Ramette, Thomas Hartke, SciSpace, Saikiran Chandha

I have never spoken publicly about a personal angel investment I've made, but I'm changing that today because I think anyone searching the scientific or technical literature regularly should be using Undermind. Google Scholar leaves me feeling like I'm sifting through low-value background information. Undermind gives me exactly what I need to make technical judgments: highly relevant papers and their overall context. Undermind.ai works by using an AI agent to perform many literature searches over several minutes to get you the information you really need, not just a list of search results. It also creates a mini-review article of all the papers it looks at and provides an LLM chatbot with knowledge of those abstracts. With an Undermind search I know: - How relevant each paper is to my question - How those papers are connected to each other - How likely it is that I'm looking at all the papers that matter for my question. Because I support dozens of startups across as many technical areas, I'm constantly asking specific questions in domains outside my core expertise. Undermind has come through at several crucial moments to save me a ton of time and effort. I'm excited to keep using it and to see what Joshua Ramette and Thomas Hartke build next.

I entered my review question.. & this tool retrieved.. ????% of the studies I included in my Systematic Review ?? YES, this tool is the MOST powerful AI literature search engine out there! It also identfied +1 new study I had missed (?????????????????? ?????????? ?????? ???????? ???????? ???? ????????????) --------------------------------------------------------- ?? ???????? ???? ???? ???????????? ?? ?????????? ?????????????? ???????? ???????? ???? ????????????? Comment '??????????' & I'll be sure to let you know when I do! --------------------------------------------------------- You can ask questions of upto 1000 characters in a single query. And it helps you get super specific with your input? ? so it can give you super relevant output ------------------------------------------ Useful find? Pass it on! ?? ???????????? ------------------------------------------ Undermind is Google for scientific research.? It finds EXACTLY what you need.. ..no matter how COMPLEX your research question is! ?? ???? ???????????? ?????????????????????? ???? ?????? ???????????????? ----------------------------------------------------- ?? Want help choosing the best AI for your research? Get my 7-hour course (20+ ????????????????) + bonuses ? ?? https://lnkd.in/gjC3ZJgj [175+ ?????? ?????????????? ???????????????? ?????? ??????????????????] And Razia + follow + ?? I test AI tools to simplify your #research & #analysis ----------------------------------------------------- #aiinresearch #scientificliterature #aisearch

Thrilled to invest in Undermind YC S24 , a search engine that can handle extremely complex questions for experts like research scientists & doctors. Many YC bio-founders are paying subscribers—and investors! Backed by Ycombinator & Pioneer Fund. If you're a researcher and want to try it out, let me know!