PenDigits with NeuralTS¶
In this tutorial, we will be training a NeuralTS agent to solve the PenDigits dataset, converted into a bandit environment. We will also use evolutionary hyperparameter optimization on a population of agents.
To complete the PenDigits environment, the agent must learn to select the best arm, or action, to take in a given context, or state.
Figure 1: Cumulative regret from training on the PenDigits dataset¶ |
Figure 2: Reward from training on the PenDigits dataset¶ |
NeuralTS (Neural Contextual Bandits with Thompson Sampling) adapts deep neural networks for both exploration and exploitation by using a posterior distribution of the reward with a neural network approximator as its mean, and neural tangent features as its variance.
For this tutorial, we will use the labelled PenDigits dataset from the UCI Machine Learning Repository.
These datasets can easily be imported and used for training with the Python package ucimlrepo, and to choose from the hundreds of
available datasets it is as simple as changing the id parameter used by fetch_uci_repo.
We can convert these labelled datasets into a bandit learning environment easily by using the agilerl.wrappers.learning.BanditEnv class.
"""This tutorial shows how to train an NeuralTS agent on the PenDigits dataset with evolutionary HPO.
Authors: Nick (https://github.com/nicku-a)
"""
import matplotlib.pyplot as plt
import numpy as np
import torch
from scipy.ndimage import gaussian_filter1d
from tqdm import trange
from ucimlrepo import fetch_ucirepo
from agilerl.components.replay_buffer import ReplayBuffer
from agilerl.hpo.mutation import Mutations
from agilerl.hpo.tournament import TournamentSelection
from agilerl.utils.utils import initialPopulation
from agilerl.wrappers.learning import BanditEnv
if __name__ == "__main__":
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
NET_CONFIG = {
"arch": "mlp", # Network architecture
"hidden_size": [128], # Actor hidden size
}
INIT_HP = {
"POPULATION_SIZE": 4, # Population size
"BATCH_SIZE": 64, # Batch size
"LR": 0.001, # Learning rate
"GAMMA": 1.0, # Scaling factor
"LAMBDA": 1.0, # Regularization factor
"REG": 0.0625, # Loss regularization factor
"LEARN_STEP": 1, # Learning frequency
# Swap image channels dimension from last to first [H, W, C] -> [C, H, W]
"CHANNELS_LAST": False,
}
# Fetch data https://archive.ics.uci.edu/
pendigits = fetch_ucirepo(id=81)
features = pendigits.data.features
targets = pendigits.data.targets
env = BanditEnv(features, targets) # Create environment
context_dim = env.context_dim
action_dim = env.arms
pop = initialPopulation(
algo="NeuralTS", # Algorithm
state_dim=context_dim, # State dimension
action_dim=action_dim, # Action dimension
one_hot=None, # One-hot encoding
net_config=NET_CONFIG, # Network configuration
INIT_HP=INIT_HP, # Initial hyperparameters
population_size=INIT_HP["POPULATION_SIZE"], # Population size
device=device,
)
field_names = ["context", "reward"]
memory = ReplayBuffer(
action_dim=action_dim, # Number of agent actions
memory_size=10000, # Max replay buffer size
field_names=field_names, # Field names to store in memory
device=device,
)
tournament = TournamentSelection(
tournament_size=2, # Tournament selection size
elitism=True, # Elitism in tournament selection
population_size=INIT_HP["POPULATION_SIZE"], # Population size
evo_step=1,
) # Evaluate using last N fitness scores
mutations = Mutations(
algo="NeuralTS", # Algorithm
no_mutation=0.4, # No mutation
architecture=0.2, # Architecture mutation
new_layer_prob=0.2, # New layer mutation
parameters=0.2, # Network parameters mutation
activation=0.2, # Activation layer mutation
rl_hp=0.2, # Learning HP mutation
rl_hp_selection=["lr", "batch_size"], # Learning HPs to choose from
mutation_sd=0.1, # Mutation strength
mutate_elite=False, # Mutate best agent in population
arch=NET_CONFIG["arch"], # Network architecture
rand_seed=1, # Random seed
device=device,
)
max_episodes = 50 # Max training episodes
max_steps = 50 # Max steps per episode
evo_epochs = 2 # Evolution frequency
evo_loop = 1 # Number of evaluation episodes
print("Training...")
regret = [[0] for _ in pop]
score = [[0] for _ in pop]
total_steps = 0
# TRAINING LOOP
for idx_epi in trange(max_episodes):
for i, agent in enumerate(pop): # Loop through population
losses = []
context = env.reset() # Reset environment at start of episode
for idx_step in range(max_steps):
# Get next action from agent
action = agent.getAction(context)
next_context, reward = env.step(action) # Act in environment
# Save experience to replay buffer
memory.save2memory(context[action], reward)
# Learn according to learning frequency
if (
memory.counter % agent.learn_step == 0
and len(memory) >= agent.batch_size
):
for _ in range(2):
experiences = memory.sample(
agent.batch_size
) # Sample replay buffer
# Learn according to agent's RL algorithm
loss = agent.learn(experiences)
losses.append(loss)
context = next_context
score[i].append(reward)
regret[i].append(regret[i][-1] + 1 - reward)
total_steps += max_steps
# Now evaluate and evolve population if necessary
if (idx_epi + 1) % evo_epochs == 0:
# Evaluate population
fitnesses = [
agent.test(
env,
swap_channels=INIT_HP["CHANNELS_LAST"],
max_steps=max_steps,
loop=evo_loop,
)
for agent in pop
]
print(f"Episode {idx_epi+1}/{max_episodes}")
print(f"Regret: {[regret[i][-1] for i in range(len(pop))]}")
# Tournament selection and population mutation
elite, pop = tournament.select(pop)
pop = mutations.mutation(pop)
# Plot the results
plt.figure()
for i, agent_regret in enumerate(regret):
plt.plot(
np.linspace(0, total_steps, len(agent_regret)),
agent_regret,
label=f"NeuralTS: Agent {i}",
)
plt.xlabel("Training Step")
plt.ylabel("Regret")
plt.legend()
plt.savefig("NeuralTS-PenDigits-regret.png")
plt.figure()
for i, agent_score in enumerate(score):
smoothed_score = gaussian_filter1d(agent_score, sigma=80)
plt.plot(
np.linspace(0, total_steps, len(smoothed_score)),
smoothed_score,
label=f"NeuralTS: Agent {i}",
)
plt.xlabel("Training Step")
plt.ylabel("Reward")
plt.legend()
plt.savefig("NeuralTS-PenDigits-reward.png")