StudyDSA logoStudyDSA

Command Palette

Search for a command to run...

Sign InSign Up
Sign Up

Data StructuresAlgorithmsBig-O NotationDesign PatternsSystem DesignMachine LearningPhysicsRoboticsAI Research

Definition
StudyDSA

Where complexity meets clarity.
By Armas Zarra.

Topics

  • Data Structures
  • Algorithms
  • Big-O Notation
  • Robotics
  • AI Research
  • Machine Learning

Practice

  • Blind 75
  • LeetCode 75
  • NeetCode 150

Legal

  • Privacy Policy
  • Terms of Service

© 2026 Armas Films LLC

IntroductionAI Research

AI Research

The craft of pushing AI forward, from reading papers to building world models.

Definition

Every few months in AI, something impossible becomes a benchmark, then a baseline, then boring. AI research is the work behind that cycle: making machines do things they couldn't do last year. It's different from applying machine learning. Instead of using known techniques to solve a business problem, you're producing new knowledge: architectures, training methods, and results nobody has published yet.


This page follows one deliberate road through the field: research practice first, then the architectures behind modern AI, then reinforcement learning, ending at world models, the systems several major labs now bet will become the physics engines of robotics.

Research Practice

Research is a craft before it's a topic. You learn to read papers in passes instead of front to back, to reproduce results before trusting or extending them, and to design experiments where an ablation actually proves something. Benchmarks get a layer of suspicion of their own: by 2026 the classic ones are saturated and leaked into training data, so knowing what a score really measures is a research skill in itself.

Modern Architectures

A handful of architectural ideas carry nearly everything modern. Attention lets every token weigh every other token at once, which scaled into transformers and then into vision. Self-supervised learning removes the labels, while diffusion and flow matching turn noise into data. Each one earned its place by beating something older, and each is worth understanding from scratch.

Reinforcement Learning

Reinforcement learning is learning by trial and error: an agent acts, the world responds, and reward is the only teacher. The split that matters most here is model-free versus model-based: whether the agent maps the world first or learns purely from outcomes. That question leads directly to imagination training, where a policy trains inside a learned simulator instead of the real world.

World Models

The capstone, and the fastest-moving corner of the field. A world model is a network that predicts what happens next, given a state and an action. It's a physics engine learned from data instead of programmed. In one recent year, V-JEPA 2, Genie 3, Dreamer 4, and NVIDIA's Cosmos all landed, and robot policies like Pi0 started turning those predictions into real-world motion. This is where the previous three tiers were heading all along.

Learning Path

Here is the whole path, tier by tier. It's ordered so each tier earns the next, ending at world models. Each topic will get its own page soon, but until then, use this as the map.

Research Practice

Before any architecture, the meta-skills: how to read, reproduce, and measure. Most failed research dies here, not in the math.

Reading Papers

The three-pass method: skim, study, then rebuild it in your head.

Coming soon!

Reproducing Results

If you can't rebuild it, you don't understand it yet.

papercode95.195.0
Coming soon!

Experiment Design

Ablations, baselines, and seeds to prove your idea is the reason.

base-attndiff
Coming soon!

Benchmarks & Evaluation

What a score really measures, and when a leaderboard lies.

Coming soon!
Modern Architectures

With the craft in place, move to the ideas themselves. Almost everything since 2017 is assembled from the pieces in this tier.

Transformers & Attention

The idea behind modern AI: every token attends to every other.

Coming soon!

Vision Transformers

Chop an image into patches and treat them like words.

Coming soon!

Self-Supervised Learning

Models that teach themselves from unlabeled data.

Coming soon!

Diffusion Models

Generates by learning to undo noise, one step at a time.

Coming soon!

Flow Matching

A straighter road from noise to data, in far fewer steps.

Coming soon!
Reinforcement Learning

The architectures above learn from data that already exists. Acting in a world that pushes back is a different problem, and this tier is trial and error made rigorous.

RL Fundamentals

The core loop: an agent acts, the reward arrives, the policy updates.

agentenv+1
Coming soon!

Model-Free vs Model-Based

React from experience, or learn a world you can plan inside.

samodel
Coming soon!

PPO & GRPO

The policy optimizers behind game agents and language models alike.

Coming soon!

Imagination Training

Let the policy practice inside the model's own dream.

world modelpolicy
Coming soon!
World Models

Everything above converges here: an architecture that predicts, trained on experience, becomes a simulator the agent can rehearse in. The bet is that learned physics engines are how robots finally generalize.

What Are World Models

Networks that predict what happens next: learned physics engines.

samodels'
Coming soon!

Latent Dynamics

Compress the world to its essentials, then predict in that space.

obs
Coming soon!

Video Prediction

Learning how the world moves from a million hours of video.

Coming soon!

Foundation World Models

Genie, Cosmos, Dreamer: one model, many worlds.

WFM
Coming soon!

Robot Policies

From predicted futures to motor commands, fifty times a second.

obspolicyrobot
Coming soon!