File size: 11,720 Bytes
c8b77b5 |
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 |
"""
Ethics Training Module for MangoMAS Local
This module implements specialized training for ethical reasoning capabilities,
adapted from the AWS backup system for local training.
"""
import json
import logging
import os
import random
from typing import Any, Dict, List
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.utils.data import Dataset
from ..core_framework import SpecializedTrainingModule, TrainingModuleConfig
logger = logging.getLogger(__name__)
class EthicsDataset(Dataset):
"""Dataset for training ethical reasoning capabilities."""
def __init__(self, data_path: str, tokenizer, max_length: int = 768):
"""
Initialize the ethics dataset.
Args:
data_path: Path to the ethics data file
tokenizer: Tokenizer for text processing
max_length: Maximum sequence length
"""
self.tokenizer = tokenizer
self.max_length = max_length
self.data = self._load_data(data_path)
logger.info(f"Loaded ethics dataset with {len(self.data)} examples")
def _load_data(self, data_path: str) -> List[Dict]:
"""Load ethics training data."""
data = []
with open(data_path, "r", encoding="utf-8") as f:
for line in f:
try:
item = json.loads(line.strip())
# Validate required fields for ethics data
if (
"scenario" in item
and "ethical_analysis" in item
and "recommendation" in item
):
data.append(item)
except json.JSONDecodeError:
continue
return data
def __len__(self):
return len(self.data)
def __getitem__(self, idx):
item = self.data[idx]
# Format the ethics prompt
prompt = f"Scenario: {item['scenario']}\nEthical Analysis: {item['ethical_analysis']}\nRecommendation: {item['recommendation']}"
# Tokenize
encoding = self.tokenizer(
prompt,
max_length=self.max_length,
padding="max_length",
truncation=True,
return_tensors="pt",
)
return {
"input_ids": encoding["input_ids"].squeeze(),
"attention_mask": encoding["attention_mask"].squeeze(),
"labels": encoding["input_ids"].squeeze(),
}
class EthicsTrainingModule(SpecializedTrainingModule):
"""Specialized training module for ethical reasoning capabilities."""
def __init__(self, config: TrainingModuleConfig, tokenizer):
"""
Initialize the ethics training module.
Args:
config: Module configuration
tokenizer: Tokenizer for text processing
"""
super().__init__(config, tokenizer)
# Initialize ethics-specific components
self.ethics_loss = nn.CrossEntropyLoss(ignore_index=-100)
self.metrics = {
"ethics_loss": 0.0,
"ethical_consistency": 0.0,
"principle_alignment": 0.0,
}
logger.info("Initialized EthicsTrainingModule")
def prepare_batch(self, batch: Dict[str, torch.Tensor]) -> Dict[str, torch.Tensor]:
"""
Prepare a batch of data for ethics training.
Args:
batch: The input batch from the dataloader
Returns:
Processed batch ready for ethics training
"""
# Move batch to device
prepared_batch = {}
for key, value in batch.items():
if isinstance(value, torch.Tensor):
prepared_batch[key] = value.to(self.device)
else:
prepared_batch[key] = value
return prepared_batch
def compute_loss(
self, student_outputs: Any, teacher_outputs: Any, batch: Dict[str, torch.Tensor]
) -> torch.Tensor:
"""
Compute the ethics-specific loss.
Args:
student_outputs: Outputs from the student model
teacher_outputs: Outputs from the teacher model
batch: The processed input batch
Returns:
Loss tensor for ethics training
"""
try:
# Extract logits from model outputs
if hasattr(student_outputs, "logits"):
student_logits = student_outputs.logits
else:
student_logits = student_outputs
if hasattr(teacher_outputs, "logits"):
teacher_logits = teacher_outputs.logits
else:
teacher_logits = teacher_outputs
# Get labels from batch
labels = batch.get("labels", batch.get("input_ids"))
# Compute cross entropy loss for ethics
shift_logits = student_logits[..., :-1, :].contiguous()
shift_labels = labels[..., 1:].contiguous()
ethics_loss = self.ethics_loss(
shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1)
)
# Add KL divergence loss between student and teacher
if teacher_logits is not None:
kl_loss = F.kl_div(
F.log_softmax(student_logits, dim=-1),
F.softmax(teacher_logits, dim=-1),
reduction="batchmean",
)
total_loss = ethics_loss + 0.1 * kl_loss
else:
total_loss = ethics_loss
# Update metrics
self.metrics["ethics_loss"] = ethics_loss.item()
return total_loss * self.loss_weight
except Exception as e:
logger.error(f"Error computing ethics loss: {e}")
# Return a small loss to avoid training failure
return torch.tensor(0.01, requires_grad=True)
def get_metrics(self) -> Dict[str, float]:
"""
Get metrics specific to ethics training.
Returns:
Dictionary of ethics metrics
"""
return self.metrics.copy()
def generate_synthetic_ethics_data(
self, output_path: str, num_samples: int = 1000
) -> None:
"""
Generate synthetic ethics training data.
Args:
output_path: Path to save the generated data
num_samples: Number of samples to generate
"""
# This is a simplified implementation based on the AWS backup
# In a full implementation, this would be much more sophisticated
ethics_templates = [
{
"scenario": "A company wants to collect user location data to improve their mapping service.",
"principles": [
"privacy",
"consent",
"transparency",
"data minimization",
],
"considerations": [
"Users should be clearly informed about what data is collected",
"Data collection should be opt-in rather than opt-out",
"Only necessary data should be collected and stored",
"Data should be properly secured and anonymized where possible",
],
"ethical_analysis": "The collection of location data raises privacy concerns, but can be ethical if done with transparency, informed consent, and data minimization practices. Users must be clearly informed about what data is collected, how it's used, and given genuine choice in the matter.",
"recommendation": "Proceed with location data collection only with explicit opt-in consent, clear privacy notices, data minimization practices, and strong security measures.",
"stance": "neutral",
},
{
"scenario": "An AI company is developing a facial recognition system to be sold to law enforcement agencies without oversight mechanisms.",
"principles": [
"privacy",
"justice",
"accountability",
"potential for discrimination",
],
"considerations": [
"Facial recognition has known bias issues across different demographics",
"Law enforcement use creates significant civil liberties concerns",
"Lack of oversight could lead to misuse and privacy violations",
"Potential chilling effect on free speech and assembly",
],
"ethical_analysis": "Deploying facial recognition in law enforcement without oversight mechanisms raises serious ethical concerns. These systems have demonstrated bias across demographic groups, potentially leading to discriminatory outcomes. Without accountability measures, there's significant risk of misuse, privacy violations, and erosion of civil liberties.",
"recommendation": "Do not deploy facial recognition systems to law enforcement without robust oversight, accuracy testing across demographics, clear usage limitations, and strong accountability mechanisms.",
"stance": "harmful",
},
{
"scenario": "A medical AI is being developed to help doctors identify potential early signs of cancer in medical images.",
"principles": [
"beneficence",
"non-maleficence",
"human oversight",
"transparency",
],
"considerations": [
"Early detection could save many lives",
"False positives could cause unnecessary stress and procedures",
"False negatives could delay critical treatment",
"System should augment rather than replace medical expertise",
],
"ethical_analysis": "A medical AI for cancer detection has significant potential benefits in improving early diagnosis and saving lives. However, it's critical that the system maintains high accuracy to minimize both false positives (causing unnecessary procedures) and false negatives (missing actual cases). The system should be designed to augment rather than replace medical professionals, with humans making final decisions.",
"recommendation": "Proceed with development with rigorous clinical validation, transparent reporting of accuracy metrics across diverse populations, clear communication about limitations, and implementation as a decision support tool rather than autonomous system.",
"stance": "beneficial",
},
]
# Generate variations
output_data = []
for _ in range(num_samples):
template = random.choice(ethics_templates)
# Create a minor variation to avoid exact duplicates
variation = template.copy()
# Add metadata
variation["metadata"] = {
"generated": True,
"timestamp": (
torch.cuda.get_device_name(0)
if torch.cuda.is_available()
else "CPU"
),
}
output_data.append(variation)
# Save to file
os.makedirs(os.path.dirname(output_path), exist_ok=True)
with open(output_path, "w", encoding="utf-8") as f:
for item in output_data:
f.write(json.dumps(item) + "\n")
logger.info(
f"Generated {len(output_data)} synthetic ethics examples at {output_path}"
)
|