[tm]

current research

Comparing Neural Representations: Theoretical vs. Deep Learning Models
july 2025 – present
active NeurIPS 2025 VSS 2026
collaborators: Sudhanshu Srivastava (UC San Diego), Ansh Soni (University of Pennsylvania)
PyTorchNumPy fMRIRSAHMAX

Using representational similarity analysis to compare classical vision models (HMAX, Gabor pyramids) against state of the art deep networks, benchmarked against human fMRI data from early visual cortex. The goal is to understand what structural properties of a model drive brain alignment, and whether simplicity can beat scale.

  • found that HMAX exhibits stronger alignment with V1 than recent hyper-complex architectures
  • accepted as extended abstract at UniReps, NeurIPS 2025 and contributed talk at VSS 2026
CNN Analysis of Background Distractors in Visual Search
feb 2025 – present
active URCA Grant
collaborators: Miguel Eckstein (UC Santa Barbara)
PyTorchNumPy scikit-learnPython

A synthetic image generation pipeline and model introspection framework studying how CNNs handle visual search in structured environments. This URCA grant funded project asks whether neural networks solve visual search the same way humans do, or whether they cheat by relying on statistical regularities in backgrounds rather than the target itself.

  • trained multiple CNN architectures to >90% accuracy; found models rely on background context rather than target features
  • presented as poster and oral talk at UCSB URCA Week 2025
Deceptive Descriptions in Visual AI Judgment
2024 – present active
collaborators: VIU Lab, UCSB
GPT-4Gemini Human StudiesNLP

Studying how humans and large language models evaluate fake expertise in image captions — whether AI systems can detect and be deceived by misleading descriptions. We curated 360 images with intentionally misleading descriptions across five knowledge domains, collected human deceptiveness ratings, and compared them to GPT-4 and Gemini-2.5 judgments.

  • follow-up study underway on decision making and trust formation under deceptive information

selected projects

StudyStack — Full Stack AI Academic Platform
january 2026
Next.jsTypeScript FastAPISupabaseVercel

An AI-powered academic workspace integrating Canvas, Gradescope, and Piazza into a single interface. An agentic AI layer reasons over course context to interpret assignments, surface deadlines, and guide study workflows. FastAPI backend with authenticated APIs and async task orchestration; Next.js frontend deployed on Vercel with Supabase for auth and persistence.

Computational Model of Dynamic Gaze Cueing
may 2024
PythonTensorFlow Drift Diffusion ModelingBayesian

Three frameworks modeling covert attention during dynamic gaze cueing: a Bayesian Ideal Observer, a 3D CNN, and a time-resolved Drift Diffusion Model calibrated against human psychophysics data. Validated against 9 observers and 10,000 trials, replicating key RT speed-up and accuracy effects.

Classify — Professor Ranking App
2023
React NativePython GCP NLPBeautifulSoup

React Native app with GCP NLP sentiment analysis and web scraping, aggregating professor reviews to help students make informed course decisions.

NFL Play Predictor
2023
KerasPython pandasscikit-learn

Predicts run, pass, punt, or kickoff from ten seasons of play by play data using a custom Keras model with extensive feature engineering.

posters & presentations

01
Hand-Engineered Image-Computable Models Can Still Outperform DNNs in V1 Similarity
Tanish Mendki, Sudhanshu Srivastava, Ansh Soni · 2026
Contributed talk · Vision Sciences Society Annual Meeting, May 2026 upcoming
02
Hand-Engineered Image-Computable Models Can Still Outperform DNNs in V1 Similarity
Tanish Mendki, Sudhanshu Srivastava, Ansh Soni · 2025
Extended abstract · UniReps Workshop, NeurIPS 2025 workshop
03
Humans and Neural Networks Learn to Use Background Structures for Efficient Visual Search
Tanish Mendki, Miguel P. Eckstein · 2025
Poster · URCA Poster Colloquium, UC Santa Barbara
04
Humans and Neural Networks Learn to Use Background Structures for Efficient Visual Search
Tanish Mendki, Miguel P. Eckstein · 2025
Oral presentation · URCA Conference, UC Santa Barbara talk