Skip to main content Início Criadores adu2021 skillxiv agent-early-experience
agent-early-experience Bridge imitation learning and experience-driven RL by collecting state-based supervision from agents' own actions. Trigger: improve agent generalization when expert demonstrations are limited and environments lack dense rewards.
Ir para a instalação Skills Marketplace Descubra e explore skills de IA criadas pela comunidade.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Mostrar detalhes do prompt Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.
npx skills add https://github.com/ADu2021/skillXiv --skill agent-early-experienceO comando permanece em uma só linha. Role horizontalmente para revisá-lo antes de copiar.
Prefere uma cópia local? Baixe os arquivos disponíveis atualmente no SkillsMP.
Baixar Zip Baixando... Mais deste repositório meaningful-kebab-case-name Convert arXiv papers into ready-to-use agent skills using category-aware extraction. First classifies the paper into one or more of 11 research categories, then applies a specialized extraction pipeline for each category — because different types of papers produce different types of usable knowledge. A single paper can yield multiple skills if it spans categories. Use this skill whenever the user wants to turn a paper into a skill, extract practical techniques from research, build a skill library from papers, convert arXiv papers into reusable agent instructions, or batch-process multiple papers into skills. Also trigger when someone asks about extracting actionable knowledge from papers, making research practical for LLM agents, or systematically converting academic contributions into structured agent capabilities.
action-quantization-behavior-cloning Establish regret bounds for behavior cloning with discretized actions combining statistical error and quantization error terms. Prove smoothness requirements for safe quantizer design, show that learning-based quantizers fail these requirements, and propose model-based augmentation to reduce error dependence from H² to H.
adaptive-lora-personalized-ranks Dynamically allocate LoRA ranks per-layer during fine-tuning instead of using fixed uniform ranks. Learn optimal rank for each layer and subject via variational framework with discretized exponential distribution, reducing memory footprint while maintaining fidelity and text-alignment.
Ocupações relacionadas SOC
Baseado na classificação ocupacional SOC
name agent-early-experience title Agent Learning via Early Experience: Implicit World Modeling and Self-Reflection version 0.0.2 engine skillxiv-v0.0.2-claude-opus-4.6 license MIT url https://arxiv.org/abs/2510.08558 keywords ["agent-learning","early-experience","world-modeling","self-reflection","imitation-learning"] description Bridge imitation learning and experience-driven RL by collecting state-based supervision from agents' own actions. Trigger: improve agent generalization when expert demonstrations are limited and environments lack dense rewards.
Agent Learning via Early Experience
Core Concept
Early Experience Learning addresses a fundamental challenge in agent training: expert demonstrations are limited and don't cover environment diversity, while many environments (websites, interactive systems) lack verifiable rewards. The approach collects supervision from the agent's own exploratory actions—using resulting states as training targets. This bridges imitation learning and RL by grounding policy in environmental dynamics through implicit world modeling and self-reflection.
The key insight: An agent's own experience (state transitions) is stronger supervision than limited expert demonstrations, and agent reflection on suboptimal actions drives improvement.
Architecture Overview
Early Experience Collection : Gather state transitions from agent's own exploratory actionsImplicit World Modeling : Learn environmental dynamics without explicit modelSelf-Reflection Mechanism : Agent evaluates its own actions and learns correctionsTwo Learning Strategies : World modeling vs. self-reflection (complementary)Foundation for RL : Provides strong initialization for downstream RL phases
Implementation Steps
1. Define the Experience Collection Protocol
Agents explore environments while recording their actions and resulting states.
class EarlyExperienceCollector :
"""
Collect state-based supervision from agent's own exploratory actions.
"""
def __init__ (self, agent_model, environment, max_steps=50 ):
self .agent = agent_model
self .env = environment
self .max_steps = max_steps
self .experiences = []
def collect_trajectory (self, task_description ):
"""
Agent explores a task, recording (state, action, result_state).
Args:
task_description: Human-readable task prompt
Returns:
Trajectory with experience tuples
"""
trajectory = {
"task" : task_description,
"states"
"actions"
"result_states"
"observations"
self
"states"
for
in
range
self
self
"observations"
self
f"Task: {task_description} \n"
f"Current observation: {observation} \n"
f"Your action: "
100
"actions"
try
self
"result_states"
if
self
"completed"
True
break
except
as
"result_states"
"error"
str
break
"num_steps"
len
"actions"
return
def
collect_batch
self, task_list, num_trajectories_per_task=3
"""
Collect multiple trajectories across task distribution.
"""
for
in
for
in
range
self
return
2. Implement Implicit World Modeling Train agent to predict result states given actions, grounding it in environment dynamics.
class ImplicitWorldModelTrainer :
"""
Train world model implicitly: predict state transitions.
"""
def __init__ (self, agent_model ):
self .agent = agent_model
def create_world_modeling_examples (self, trajectories ):
"""
Convert trajectories into (state, action, predicted_state) tuples.
"""
training_examples = []
for traj in trajectories:
for step_idx in range (len (traj["actions" ])):
current_state = traj["states" ][step_idx]
action = traj["actions" ][step_idx]
result_state = traj["result_states" ][step_idx]
example = {
"input" : f"State: {format_state(current_state)} \n"
f"Action: {action} \n"
f"Resulting state: " ,
"target" : format_state(result_state)
}
training_examples.append(example)
return training_examples
def train_world_model (self, examples, num_epochs=5 ):
"""
Fine-tune agent to predict state transitions.
This implicitly teaches environmental dynamics.
"""
optimizer = torch.optim.Adam(self .agent.parameters(), lr=1e-5 )
for epoch in range (num_epochs):
epoch_loss = 0
for example in examples:
output = self .agent.generate(
example["input" ],
max_tokens=200
)
loss = compute_state_matching_loss(
output,
example["target" ]
)
epoch_loss += loss.item()
optimizer.zero_grad()
loss.backward()
optimizer.step()
avg_loss = epoch_loss / len (examples)
print (f"Epoch {epoch} : world model loss = {avg_loss:.4 f} " )
return self .agent
3. Implement Self-Reflection Mechanism Agents evaluate their own actions and learn from mistakes without external feedback.
class SelfReflectionTrainer :
"""
Train agents to reflect on suboptimal actions and improve.
"""
def __init__ (self, agent_model ):
self .agent = agent_model
def create_reflection_examples (self, trajectories, success_threshold=0.8 ):
"""
Identify suboptimal actions and create reflection examples.
"""
reflection_examples = []
for traj in trajectories:
task_success = traj.get("completed" , False )
num_steps = traj["num_steps" ]
efficiency = 1.0 / (num_steps + 1 )
success_score = 1.0 if task_success else 0.0
overall_score = 0.7 * success_score + 0.3 * efficiency
if overall_score < success_threshold:
for step_idx in range (len (traj["actions" ])):
action = traj["actions" ][step_idx]
state = traj["states" ][step_idx]
result = traj["result_states" ][step_idx]
progress = self .evaluate_action_progress(
action,
state,
result,
traj["task" ]
)
if progress < 0.5 :
reflection_example = {
"input" : f"Task: {traj['task' ]} \n"
f"Current state: {format_state(state)} \n"
f"Action taken: {action} \n"
f"Result: {format_state(result)} \n"
f"Reflection (what went wrong): " ,
"target" : generate_reflection(action, result)
}
reflection_examples.append(reflection_example)
return reflection_examples
def train_self_reflection (self, examples, num_epochs=5 ):
"""
Train agent to articulate what went wrong with its actions.
"""
optimizer = torch.optim.Adam(self .agent.parameters(), lr=1e-5 )
for epoch in range (num_epochs):
epoch_loss = 0
for example in examples:
output = self .agent.generate(
example["input" ],
max_tokens=100
)
loss = compute_reflection_loss(
output,
example["target" ]
)
epoch_loss += loss.item()
optimizer.zero_grad()
loss.backward()
optimizer.step()
avg_loss = epoch_loss / len (examples)
print (f"Epoch {epoch} : reflection loss = {avg_loss:.4 f} " )
return self .agent
def evaluate_action_progress (self, action, state, result, task ):
"""
Measure if action moved agent closer to task goal.
"""
state_embedding = encode_state(state)
result_embedding = encode_state(result)
task_embedding = encode_task(task)
similarity_improvement = (
cosine_similarity(result_embedding, task_embedding) -
cosine_similarity(state_embedding, task_embedding)
)
return max (0.0 , similarity_improvement)
4. Combined Early Experience Training Integrate world modeling and self-reflection into coherent training pipeline.
def train_with_early_experience (agent, environment, task_list, config ):
"""
Full early experience training pipeline.
"""
print ("Phase 1: Collecting early experiences" )
collector = EarlyExperienceCollector(agent, environment)
trajectories = collector.collect_batch(
task_list,
num_trajectories_per_task=3
)
print (f"Collected {len (trajectories)} trajectories" )
print ("\nPhase 2: Implicit world modeling" )
world_trainer = ImplicitWorldModelTrainer(agent)
wm_examples = world_trainer.create_world_modeling_examples(trajectories)
agent = world_trainer.train_world_model(wm_examples, num_epochs=5 )
print ("\nPhase 3: Self-reflection learning" )
reflection_trainer = SelfReflectionTrainer(agent)
ref_examples = reflection_trainer.create_reflection_examples(trajectories)
agent = reflection_trainer.train_self_reflection(
ref_examples,
num_epochs=5
)
return agent
5. Evaluation and RL Foundation Assess early experience training and prepare for downstream RL.
def evaluate_early_experience_agent (agent, benchmark_tasks ):
"""
Test agent on diverse tasks post-early-experience training.
"""
success_rates = {}
generalization_scores = []
for task in benchmark_tasks:
successes = 0
trials = 5
for _ in range (trials):
trajectory = collect_single_trajectory(agent, task)
if trajectory.get("completed" ):
successes += 1
task_success_rate = successes / trials * 100
success_rates[task] = task_success_rate
generalization = measure_transfer_learning(agent, task)
generalization_scores.append(generalization)
print (f"Average success rate: {np.mean(list (success_rates.values())):.1 f} %" )
print (f"Generalization: {np.mean(generalization_scores):.2 f} " )
return {
"success_rates" : success_rates,
"generalization" : np.mean(generalization_scores)
}
Practical Guidance
Trajectories per task : 3-5 (balance diversity vs. compute)Max steps per trajectory : 20-50 (depends on task complexity)World modeling epochs : 5 (avoid overfitting)Self-reflection epochs : 5Learning rate : 1e-5 (conservative for LLMs)
Training agents for web environments, APIs, or interactive tasks
Limited expert demonstrations available
Want to improve generalization with self-supervised learning
As initialization before downstream RL
High-reward environments (standard RLHF sufficient)
Agents require immediate deployment (training overhead)
Tasks with no natural state representation
Single-turn interactions without state progression
Reference 