Spaces:
Runtime error
Runtime error
File size: 15,138 Bytes
1d75522 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 |
"""Enhanced learning mechanisms for reasoning strategies."""
import logging
from typing import Dict, Any, List, Optional, Set, Union, Type, Tuple
import json
from dataclasses import dataclass, field
from enum import Enum
from datetime import datetime
import numpy as np
from collections import defaultdict
@dataclass
class LearningEvent:
"""Event for strategy learning."""
strategy_type: str
event_type: str
data: Dict[str, Any]
outcome: Optional[float]
timestamp: datetime = field(default_factory=datetime.now)
class LearningMode(Enum):
"""Types of learning modes."""
SUPERVISED = "supervised"
REINFORCEMENT = "reinforcement"
ACTIVE = "active"
TRANSFER = "transfer"
META = "meta"
ENSEMBLE = "ensemble"
@dataclass
class LearningState:
"""State for learning process."""
mode: LearningMode
parameters: Dict[str, Any]
history: List[LearningEvent]
metrics: Dict[str, float]
metadata: Dict[str, Any] = field(default_factory=dict)
class EnhancedLearningManager:
"""
Advanced learning manager that:
1. Implements multiple learning modes
2. Tracks learning progress
3. Adapts learning parameters
4. Optimizes strategy performance
5. Transfers knowledge between strategies
"""
def __init__(self,
learning_rate: float = 0.1,
exploration_rate: float = 0.2,
memory_size: int = 1000):
self.learning_rate = learning_rate
self.exploration_rate = exploration_rate
self.memory_size = memory_size
# Learning states
self.states: Dict[str, LearningState] = {}
# Performance tracking
self.performance_history: List[Dict[str, Any]] = []
self.strategy_metrics: Dict[str, List[float]] = defaultdict(list)
# Knowledge transfer
self.knowledge_base: Dict[str, Any] = {}
self.transfer_history: List[Dict[str, Any]] = []
async def learn(self,
strategy_type: str,
event: LearningEvent,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Learn from strategy execution event."""
try:
# Initialize or get learning state
state = self._get_learning_state(strategy_type)
# Select learning mode
mode = await self._select_learning_mode(event, state, context)
# Execute learning
if mode == LearningMode.SUPERVISED:
result = await self._supervised_learning(event, state, context)
elif mode == LearningMode.REINFORCEMENT:
result = await self._reinforcement_learning(event, state, context)
elif mode == LearningMode.ACTIVE:
result = await self._active_learning(event, state, context)
elif mode == LearningMode.TRANSFER:
result = await self._transfer_learning(event, state, context)
elif mode == LearningMode.META:
result = await self._meta_learning(event, state, context)
elif mode == LearningMode.ENSEMBLE:
result = await self._ensemble_learning(event, state, context)
else:
raise ValueError(f"Unsupported learning mode: {mode}")
# Update state
self._update_learning_state(state, result)
# Record performance
self._record_performance(strategy_type, result)
return result
except Exception as e:
logging.error(f"Error in learning: {str(e)}")
return {
"success": False,
"error": str(e),
"mode": mode.value if 'mode' in locals() else None
}
async def _supervised_learning(self,
event: LearningEvent,
state: LearningState,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Implement supervised learning."""
# Extract features and labels
features = await self._extract_features(event.data, context)
labels = event.outcome if event.outcome is not None else 0.0
# Train model
model_update = await self._update_model(features, labels, state, context)
# Validate performance
validation = await self._validate_model(model_update, state, context)
return {
"success": True,
"mode": LearningMode.SUPERVISED.value,
"model_update": model_update,
"validation": validation,
"metrics": {
"accuracy": validation.get("accuracy", 0.0),
"loss": validation.get("loss", 0.0)
}
}
async def _reinforcement_learning(self,
event: LearningEvent,
state: LearningState,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Implement reinforcement learning."""
# Extract state and action
current_state = await self._extract_state(event.data, context)
action = event.data.get("action")
reward = event.outcome if event.outcome is not None else 0.0
# Update policy
policy_update = await self._update_policy(
current_state, action, reward, state, context)
# Optimize value function
value_update = await self._update_value_function(
current_state, reward, state, context)
return {
"success": True,
"mode": LearningMode.REINFORCEMENT.value,
"policy_update": policy_update,
"value_update": value_update,
"metrics": {
"reward": reward,
"value_error": value_update.get("error", 0.0)
}
}
async def _active_learning(self,
event: LearningEvent,
state: LearningState,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Implement active learning."""
# Query selection
query = await self._select_query(event.data, state, context)
# Get feedback
feedback = await self._get_feedback(query, context)
# Update model
model_update = await self._update_model_active(
query, feedback, state, context)
return {
"success": True,
"mode": LearningMode.ACTIVE.value,
"query": query,
"feedback": feedback,
"model_update": model_update,
"metrics": {
"uncertainty": query.get("uncertainty", 0.0),
"feedback_quality": feedback.get("quality", 0.0)
}
}
async def _transfer_learning(self,
event: LearningEvent,
state: LearningState,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Implement transfer learning."""
# Source task selection
source_task = await self._select_source_task(event.data, state, context)
# Knowledge extraction
knowledge = await self._extract_knowledge(source_task, context)
# Transfer adaptation
adaptation = await self._adapt_knowledge(
knowledge, event.data, state, context)
# Apply transfer
transfer = await self._apply_transfer(adaptation, state, context)
return {
"success": True,
"mode": LearningMode.TRANSFER.value,
"source_task": source_task,
"knowledge": knowledge,
"adaptation": adaptation,
"transfer": transfer,
"metrics": {
"transfer_efficiency": transfer.get("efficiency", 0.0),
"adaptation_quality": adaptation.get("quality", 0.0)
}
}
async def _meta_learning(self,
event: LearningEvent,
state: LearningState,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Implement meta-learning."""
# Task characterization
task_char = await self._characterize_task(event.data, context)
# Strategy selection
strategy = await self._select_strategy(task_char, state, context)
# Parameter optimization
optimization = await self._optimize_parameters(
strategy, task_char, state, context)
# Apply meta-learning
meta_update = await self._apply_meta_learning(
optimization, state, context)
return {
"success": True,
"mode": LearningMode.META.value,
"task_characterization": task_char,
"strategy": strategy,
"optimization": optimization,
"meta_update": meta_update,
"metrics": {
"strategy_fit": strategy.get("fit_score", 0.0),
"optimization_improvement": optimization.get("improvement", 0.0)
}
}
async def _ensemble_learning(self,
event: LearningEvent,
state: LearningState,
context: Dict[str, Any]) -> Dict[str, Any]:
"""Implement ensemble learning."""
# Member selection
members = await self._select_members(event.data, state, context)
# Weight optimization
weights = await self._optimize_weights(members, state, context)
# Combine predictions
combination = await self._combine_predictions(
members, weights, event.data, context)
return {
"success": True,
"mode": LearningMode.ENSEMBLE.value,
"members": members,
"weights": weights,
"combination": combination,
"metrics": {
"ensemble_diversity": weights.get("diversity", 0.0),
"combination_strength": combination.get("strength", 0.0)
}
}
def _get_learning_state(self, strategy_type: str) -> LearningState:
"""Get or initialize learning state for strategy."""
if strategy_type not in self.states:
self.states[strategy_type] = LearningState(
mode=LearningMode.SUPERVISED,
parameters={
"learning_rate": self.learning_rate,
"exploration_rate": self.exploration_rate
},
history=[],
metrics={}
)
return self.states[strategy_type]
def _update_learning_state(self, state: LearningState, result: Dict[str, Any]):
"""Update learning state with result."""
# Update history
state.history.append(LearningEvent(
strategy_type=result.get("strategy_type", "unknown"),
event_type="learning_update",
data=result,
outcome=result.get("metrics", {}).get("accuracy", 0.0),
timestamp=datetime.now()
))
# Update metrics
for metric, value in result.get("metrics", {}).items():
if metric in state.metrics:
state.metrics[metric] = (
0.9 * state.metrics[metric] + 0.1 * value # Exponential moving average
)
else:
state.metrics[metric] = value
# Adapt parameters
self._adapt_parameters(state, result)
def _record_performance(self, strategy_type: str, result: Dict[str, Any]):
"""Record learning performance."""
self.performance_history.append({
"timestamp": datetime.now().isoformat(),
"strategy_type": strategy_type,
"mode": result.get("mode"),
"metrics": result.get("metrics", {}),
"success": result.get("success", False)
})
# Update strategy metrics
for metric, value in result.get("metrics", {}).items():
self.strategy_metrics[f"{strategy_type}_{metric}"].append(value)
# Maintain memory size
if len(self.performance_history) > self.memory_size:
self.performance_history = self.performance_history[-self.memory_size:]
def _adapt_parameters(self, state: LearningState, result: Dict[str, Any]):
"""Adapt learning parameters based on performance."""
# Adapt learning rate
if "accuracy" in result.get("metrics", {}):
accuracy = result["metrics"]["accuracy"]
if accuracy > 0.8:
state.parameters["learning_rate"] *= 0.95 # Decrease if performing well
elif accuracy < 0.6:
state.parameters["learning_rate"] *= 1.05 # Increase if performing poorly
# Adapt exploration rate
if "reward" in result.get("metrics", {}):
reward = result["metrics"]["reward"]
if reward > 0:
state.parameters["exploration_rate"] *= 0.95 # Decrease if getting rewards
else:
state.parameters["exploration_rate"] *= 1.05 # Increase if not getting rewards
# Clip parameters to reasonable ranges
state.parameters["learning_rate"] = np.clip(
state.parameters["learning_rate"], 0.001, 0.5)
state.parameters["exploration_rate"] = np.clip(
state.parameters["exploration_rate"], 0.01, 0.5)
def get_performance_metrics(self) -> Dict[str, Any]:
"""Get comprehensive performance metrics."""
return {
"learning_states": {
strategy_type: {
"mode": state.mode.value,
"parameters": state.parameters,
"metrics": state.metrics
}
for strategy_type, state in self.states.items()
},
"strategy_performance": {
metric: {
"mean": np.mean(values) if values else 0.0,
"std": np.std(values) if values else 0.0,
"min": min(values) if values else 0.0,
"max": max(values) if values else 0.0
}
for metric, values in self.strategy_metrics.items()
},
"transfer_metrics": {
"total_transfers": len(self.transfer_history),
"success_rate": sum(1 for t in self.transfer_history if t.get("success", False)) / len(self.transfer_history) if self.transfer_history else 0
}
}
def clear_history(self):
"""Clear learning history and reset states."""
self.states.clear()
self.performance_history.clear()
self.strategy_metrics.clear()
self.transfer_history.clear()
|