Innovation Group

AIfSR Innovation Group

• Provides researchers in natural sciences an easy free way to explore the usefulness of AI in their research

• Lead by experienced students. Work executed by students with expertise

• Engages students in authentic research education program with a consulting-type arrangement

  • More than 100 students participated to date

  • More than 20 projects and collaborators inside and outside of NYU and counting

  • Comprises of members enrolled in various ways

    • A core: AIfSR VIP group at Tandon (all across the university: Tandon, FAS, CDS, Courant, Abu Dhabi, Gallatin, NYU Shanghai, and more)

    • Ph.D. students

    • Members of AIfSR who already graduated from NYU but still continue their work at AIfSR

• Based on the original methodology

  • Scalable authentic research education framework (arxiv.org) 

  • Watch on YouTube

Good for researchers in academia and/or industry who

• Work in one of the natural sciences

• Would like to explore if AI can be useful for their data and/or research pipeline

• Would like to explore additional avenues in how AI can help with their unique tasks

Arrangement

• We expect you to share something of unique value, like an original dataset, original vision, or something similar that would make this project valuable for students.

• We manage our teams ourselves, and we don't "send students to you to become part of collaborator's research group"

• Data privacy: we don't share data provided by collaborators with people outside of AIfSR and make clear to our members they can't do that either. Unless specifically permitted by the collaborator

• Intellectual property: case-by-case. The AIfSR program is a mix of education, research, and service activities. As such, intellectual property questions will be addressed on a case-by-case basis

On our side

• You would be working with one of the teams of the AIfSR group we manage

• A team of students is formed out of students with complementary skills capable of understanding research processes and data and developing usable solutions

• We manage and educate our teams ourselves

• Each project goes through three stages: Formulate Project, Develop Solution, Deliver Solution

• Student teams have the freedom to drive projects on their own in a direction they believe will bring the most benefits for collaborators

Project Examples 

We have worked on or are currently working on the following projects with collaborators from NYU FAS, NYU AD, NYU Langone, Groningen University, Adámas Nanotechnologies, Semantic Scholar, and more

Biology

• Vaccination response prediction​

• Sleep patterns classification in Electroencephalography data

• Optical microscope images segmentation for determination of cells' shapes

• Image-based prediction for worms embryonic lethality ​

• Classifying disease (malaria or not) based on symptoms

• Genetics: protein sequences analysis with Large Language Model (LLM)

Chemistry

• Determining the electric current in batteries from the magnetic field around them

• Scoring and sampling for molecular crystal ranking

• Cancer Detection through Mass Spectrometry Imaging using ML​

Physics

• Focused Ion Beam (FIB) beam profile determination based on etching profiles

• Analysis of motion trajectories of particles under a microscope

• ML-enhanced analysis of spectral data collected for micro-diamonds

• Detection and classification of pests' presence based on acoustics data

• AI-enabled analysis of the movement of antiferromagnetic domain walls over time

Psychology

• Generative AI model for morphemes and connection to Magnetoencephalography data (Neurolinguistics of Morphology)

• Use of LLMs to label generic gender categories text

Scientific Software

• Automatic annotation generation for scientific figures

• Addressing challenges of automatic parsing of scientific papers PDFs

• Improving benchmarking for cloud coverage detection software with a performance evaluation on HPC