Jakov Ivan

Blog Posts

  1. Shopping Around for MLLMs: Which Model Best Transcribes Handwritten Text?

    Partake in a delightful feast as we shop around for the best MLLM for handwritten text recognition! I evaluated GPT-4o (August and May versions), GPT-4o mini, Claude Sonnet 3.5, Opus, and Haiku. Which model ruled them all? And which MLLM exhibited a creepy behavior? 😱 Read to find out!

  2. The Prompt-tastic Art for Better Handwritten Text Recognition

    Let's master the art of prompt engineering! In this post, I explore how different prompting tactics can dramatically enhance GPT-4o-Mini's ability to transcribe my handwritten journal notes, reducing error rates and increasing consistency.

  3. Evaluating GPT-4o-Mini for Handwritten Text Recognition

    Here I'll detail my process of evaluating GPT-4o-Mini's capability of transcribing handwritten text from my journal notes, shedding light to some unexpected results!

  4. Digitizing Memories: Using GPT-4o to Transcribe My Handwritten Journal Notes

    Discover how GPT-4o transcribed my handwritten journal notes into digital format, making it easier to organize and reflect on my thoughts.

  5. Making Math Fun: The Chocolate Chili Game

    Dive into a world where chocolates, chilis, and algorithms collide in a playful journey through programming basics. Discover how I turned a classroom game into 'Dodge n Fudge'—and yes, you can play too!

  6. Advancing Computational Psychiatry by Linking Theory to Practice

    This is my first blog post entry to the series which I’d call ‘The Stochastic Brain Review’. In this series, I will review research articles which pique my interest, get into their core, and explore some of their promising implications.

  7. Ready, Integrate, Fire! Part 2: Simulating a simple neuronal model

    Let's simulate a leaky integrate-and-fire neuron model with static input current using Numpy! We will also verify the formula for theoretical firing rate we derived in the last blog post.

  8. Ready, Integrate, Fire! Part 1: Solving a simple neuronal model

    For my first blog post, I will find the analytical solution of a leaky integrate-and-fire (LIF) model of a neuron under a constant injected current. Too much to handle? Read more to find out! (This is Part 1 of a series of two posts on LIF model)