← BITS subjects

// BITS · DEEP REINFORCEMENT LEARNING

Reinforcement Learning.

Learning by trial and reward — the agent–environment loop, multi-armed bandits and the explore/exploit dilemma, Markov decision processes, returns and value functions, the Bellman equations, dynamic programming, and Monte Carlo methods. This vault holds the same six resources as every subject: cheatsheet, slides explained, question bank, formula sheet, book explained, and references. Covers CS1–CS7 (mid-sem), textbook Sutton & Barto 2e.

#rl#agents#mdp#bellman#monte-carlo

WHAT'S IN THIS VAULT

★ solved paper

Mid-sem Solved Paper

The full mid-sem paper worked end-to-end — collapsible Q&A — plus a makeup-exam study guide on the same concepts.

01 · cheatsheet

Cheatsheet live

Dense one-glance reference — vocabulary, the Bellman equations, ε-greedy, value & policy iteration, Monte Carlo skeletons. CS1–CS7.

02 · slides explained

Slides Explained live

Every lecture slide unpacked in plain easy words — the why behind bandits, MDPs, Bellman, DP and Monte Carlo. CS1–CS7 (mid-sem).

03 · question bank

Question Bank live

Exam-style questions pulled straight from the slides. Questions only — no answers — plus the worked exercises. CS1–CS7.

04 · formula sheet

Formula Sheet live

Every equation for the mid-sem — action-value update, return, model dynamics, Bellman expectation & optimality, DP and MC — each symbol defined.

05 · book explained

Book Explained live

Sutton & Barto Ch 1–5, chapter by chapter, in easy language with diagrams, worked examples, gotchas, and a concept cheatsheet.

06 · references

References live

Textbook (Sutton & Barto), David Silver's course, Berkeley CS285, and the key papers — DQN, A3C, PPO, AlphaGo — worth keeping.


01 · CHEATSHEET

Cheatsheet

A dense, scannable card: the vocabulary (agent, environment, state, action, reward, return, policy, value), the core equations (action-value update, Bellman expectation/optimality, value & policy iteration, Monte Carlo), and one-line algorithm skeletons — to cram from before the mid-sem.

Open RL cheatsheet → CS1–CS7, one card

02 · SLIDES EXPLAINED

Slides Explained

Each lecture deck, slide by slide, rewritten in plain easy words — full concepts not bullet points. The intuition behind the agent–environment loop, multi-armed bandits and explore/exploit, Markov decision processes, returns and value functions, the Bellman equations, dynamic programming, and Monte Carlo methods, with the why spelled out. Covers CS1–CS7 (mid-sem).

Open RL slides explained → CS1–CS7 all live

03 · QUESTION BANK

Question Bank

Questions extracted directly from the CS1–CS7 slides — questions only, no answers — grouped by lecture, for active recall and exam practice. Work them cold, then check yourself against the slides explained.

Open RL question bank → CS1–CS7, no answers

04 · FORMULA SHEET

Formula Sheet

Every equation for the mid-sem — the action-value estimate and incremental update, the return and discounted return, model dynamics, the state- and action-value functions, the Bellman expectation and optimality equations, the policy/value-iteration updates and the Monte Carlo estimate — each symbol defined.

Open RL formula sheet → every symbol defined

05 · BOOK EXPLAINED

Book Explained

Companion notes for Richard S. Sutton & Andrew G. Barto — Reinforcement Learning: An Introduction (2nd ed., MIT Press), chapters 1–5, in plain language with diagrams and worked examples.

Open RL book explained → Sutton & Barto Ch1–5, live

06 · REFERENCES

References

The core textbook (Sutton & Barto), the foundational papers (DQN, A3C, PPO, AlphaGo), lecture series (David Silver's UCL course, Berkeley CS285), and extra reads worth keeping.

Open RL references → books, papers, courses