For years, Artificial Intelligence (AI) and Machine Learning (ML) remained largely within computer science departments and their immediate collaborators. Then, in late 2022, generative AI became widely accessible and suddenly, researchers across all disciplines had powerful new tools at their disposal. Experts proclaimed that scientific research would never be the same again. But the reality has proven more complex. Generative AI can do remarkable things, but it also generates wrong information with complete confidence. Some researchers swear by AI's help in their academic research, while others swear at it. Meanwhile, there's an overwhelming number of tools with comparable and contrasting functions and subscription models, ranging from free to cheap to hundreds of dollars monthly, with various levels of limitations. Cost-effective approaches exist in many cases, but they usually require technical setup that feel daunting. Most online guides either oversell AI's magic or assume you're a developer. This leaves a significant gap for researchers who want to use AI effectively and mindfully, without changing what they work on or becoming AI experts. Our workshop is for such researchers.
The aim of the Workshop is to familiarize researchers with Generative AI solutions in the following contexts:
Using AI as a bouncing board to brainstorm your ideas
Searching the literature for relevant papers
Assimilating the papers and taking relevant notes
Using the notes to further sharpen your arguments
Creating first drafts for your writings / presentations / posters
Issues (both practical and ethical) in the above contexts.
With the above objectives in mind, the three days will be structured as follows:
Day 1
Familiarizing oneself with the various tools that are already available in the market for the various purposes mentioned above.
Deep-dive into some of the tools that we believe are of particular importance, and familiarize the participants with their pros and cons. We will also make practical suggestions in the light of what has worked for us and what has not.
Day 2
Quick introduction to LLMs and how they work.
Quick survey of some of the popular platforms like ChatGPT, Claude, Gemini etc, with an understanding of how much they cost and their respective strengths and weaknesses.
Understanding APIs and how they can be used to significantly reduce the costs of using LLMs. Understanding how to setup APIs. Do's and Don'ts.
Setting up a personal system for running multiple LLMs from ones own computer.
Using the same system to run LLMs locally so that no data ever leaves your computer.
Understanding how to talk to LLMs. What works and what does not.
Day 3
Making Assistants to reliably improve the quality of outputs from LLMs.
Take a number of Assistants provided by us and use / tweak them to improve the quality of your output.
Use Google Colab to run open-source models for free or at a very cheap price.
Mini-project (in groups) to put together everything that you have learnt to solve a research problem end-to-end. This will include brainstorming a problem, coming up with a solution, creating a short report and making a short presentation.
The lectures will be brief. The participants will spend a lot of time working on their laptops either on their own or in small groups. There will be a number of student mentors (mostly from the instructor's group) who will be helping the participants throughout the workshop.
All participants must bring their own laptop for this course. This will enable them to try out the various things that we will provide them and leave with a setup that they can then use in their own research. Please make sure that you have a modern OS (at least Windows 10, macOS 15, Ubuntu 24.04) with at least 8 GB of RAM (16 GB or more will be better though).
Specifications for running local models: Which local model you will be able to run will depend on the configuration of your laptop. Better laptop specs -> Bigger Models -> Generally, higher quality output. For running local LLMs, on top of the modern OS (see point 2) you should have at least 16GB RAM and 10 GB+ space on your laptop. If you also have a dedicated graphics card (at least 4GB), then the experience will be smoother. Else, things will be a bit slow. Note that running local LLMs is only one part of the workshop. Therefore, even if your laptop does not meet the specs in this point, there is still a lot for you to gain from this workshop.
All participants must also bring good quality headphones. We are going to provide instruction videos on how to set up various things on the laptops, and you will need to use these headphones to listen and follow along. Headphones are generally better for long-term listening than ear buds.
We will provide each participants with sufficient individual API credits during the workshop to try out everything that we will show. However, if the participants abuse the API, and exhaust the provided API limits, then we will not be able to provide any further API credits. The participants will have to bear the cost of any further API credits during the workshop on their own. Also the API access will cease once the workshop ends.
If you are tired of wondering whether AI could actually help your research but don't know where to start beyond the basic chatbot experience, this intensive hands-on workshop will get you there. This workshop is for you if you want to:
Figure out when to trust AI and when not to, because knowing the limitations is just as important as knowing the possibilities.
Master the art of talking to AI to get dramatically better results (hint: it's not about being polite!).
Compare research-focused AI tools and their capabilities, and find the right mix of free and paid options that actually make sense for your budget and research needs.
Build your own AI assistants that work exactly how you need them to, without having to learn programming.
Walk away with a toolkit that is ready to use. We are talking actual AI workflows installed and configured on your laptop, not just theoretical knowledge.
Handle the tricky ethical stuff responsibly, from publication guidelines to data privacy, so you can use AI confidently without stepping on landmines.
In short, this workshop will let you move beyond basic ChatGPT conversations and actually leverage AI as a serious research tool.
This Workshop Will NOT
Turn you into an AI expert or domain specialist overnight.
Teach you to find research that doesn't exist.
Enable you to produce publication-ready work without verification.
Guarantee immediate research breakthroughs or productivity gains: The workshop will be a beginning, but do not expect overnight magic.
Replace your need for traditional research skills
This workshop will teach you to use AI as a powerful research assistant, not as a substitute for scholarly expertise