tutorials/attention_is_all_you_need/transformer-extended.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Critical Exploration of Transformer Models\n",
    "\n",
    "In this tutorial, we're not just using transformers; we're delving into their inner workings to understand and influence their behavior. We will aim to think critically about transformers beyond their immediate utility in tasks like classification or translation. \n",
    "\n",
    "### What We'll Explore:\n",
    "\n",
    "1. **Adversarial Inputs**: Learn to craft inputs that challenge and probe the limits of language models.\n",
    "2. **Attention Visualization**: Gain insights into how transformers focus on different parts of the input and output.\n",
    "3. **Fine-Tuning with LoRA**: Understand how to implement Low-Rank Adaptation (LoRA) for model finetuning.\n",
    "4. **Bias Detection**: Investigate inherent biases in model responses.\n",
    "\n",
    "**Practical:**: We've compiled all classes from our practical sessions into `utils.py` and `models.py`, with minor adjustments for detailed analyses like extracting attention scores. We'll use a model trained in the 'Attention is All You Need' practical session, so ensure your initializing parameters are present in `models` directly.\n",
    "\n",
    "**Note:** This session is designed to be lightweiggt. We'll explore a few key methods to probe and augment transformers without delving into complex research papers or methodologies. Consider this a starting point - the potential applications and explorations in this field are vast. Should any topic pique your interest or seem pertinent to your work, don't hesitate to reach out to your tutor for more in-depth resources and references."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda\n"
     ]
    }
   ],
   "source": [
    "from utils import *\n",
    "from models import *\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model directory to load the best model\n",
    "model_dir = pathlib.Path(\"./models\").resolve()\n",
    "\n",
    "# load the best model\n",
    "N_ENCODERS = 3\n",
    "N_DECODERS = 3\n",
    "EMB_SIZE = 128\n",
    "N_HEADS = 8\n",
    "FFN_HID_DIM = 256\n",
    "model = Transformer(SRC_N_TOKENS, TGT_N_TOKENS, N_ENCODERS, N_DECODERS, EMB_SIZE, N_HEADS, FFN_HID_DIM)\n",
    "model.load_state_dict(torch.load(model_dir / \"best.ckpt\"))\n",
    "model = model.to(device) \n",
    "\n",
    "loss_fn = LabelSmoothingLoss(TGT_N_TOKENS, 0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Crafting Adversarial Inputs for Transformers\n",
    "\n",
    "Adversarial inputs are crucial for testing and improving the robustness of transformer models. In this tutorial, we'll explore two primary methods to generate these inputs:\n",
    "\n",
    "#### 1. **Gradient Propagation**\n",
    "\n",
    "We'll employ the Fast Gradient Sign Method (FGSM) (Goodfellow et al. 2014) to create adversarial inputs.\n",
    " - Compute the loss function for a specific input.\n",
    " - Calculate the gradient of this loss.\n",
    " - Apply the negative of this gradient to perturb the input subtly.\n",
    "\n",
    "The perturbed inputs, while slightly altered, shouldn't change the target, yet they result in increased loss.\n",
    "\n",
    "**Note**: This method works at the embeddings level, not directly with word-level inputs.\n",
    "\n",
    "#### 2. **Manual Crafting Using Synonyms**\n",
    "\n",
    "We manually craft inputs by substituting words with their synonyms. The output should remain unchanged with these substitutions. Any input that causes a change in output under this substitution is considered an adversarial input.\n",
    "\n",
    "Through these techniques, we'll gain a deeper understanding of the vulnerabilities and resilience of transformer models, aiding in their development and refinement.\n",
    "\n",
    "[[Goodfellow et al. 2014](!https://arxiv.org/abs/1412.6572)] Explaining and Harnessing Adversarial Examples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Gradient Propagation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss before perturbation:  2.429346\n",
      "loss after perturbation:  2.502872\n",
      "percent increase in the loss: 3.03%\n"
     ]
    }
   ],
   "source": [
    "src, tgt = get_sample_batch('train', 1)\n",
    "\n",
    "# Copy-pasted from `process` \n",
    "src = src.to(device)\n",
    "tgt = tgt.to(device)\n",
    "\n",
    "# ignore last token in the tgt because at the time of predicting the last token, available inputs will be everything before that token\n",
    "tgt_input = tgt[:, :-1]\n",
    "\n",
    "# create masks\n",
    "src_padding_mask, tgt_mask = create_masks(src, tgt_input)\n",
    "src_padding_mask = src_padding_mask.to(device)\n",
    "tgt_mask = tgt_mask.to(device)\n",
    "\n",
    "# Creating input embeddings from the source and detaching to avoid tracking history\n",
    "input_embeddings = model.src_embeddings(src).detach()\n",
    "# Enabling gradient calculation for the input embeddings\n",
    "input_embeddings.requires_grad = True\n",
    "\n",
    "logits, attn_scores = model(src, tgt_input, src_padding_mask, tgt_mask, input_embeddings)\n",
    "\n",
    "# target predictions are made for 1st token onwards. 0th token is assumed to be <BOS> by design of the model. Read no-peak masking (above) for more details.\n",
    "tgt_out = tgt[:, 1:]\n",
    "\n",
    "# compute loss\n",
    "loss = loss_fn(logits, tgt_out)\n",
    "print(f'loss before perturbation: {loss: 0.6f}')\n",
    "\n",
    "# Resetting gradients to zero before backward pass\n",
    "model.zero_grad()\n",
    "loss.backward()\n",
    "\n",
    "# Retrieving the gradient of the input\n",
    "### YOUR CODE HERE (extract the gradients of input_embeddings)\n",
    "# input_gradients = \n",
    "\n",
    "# Applying Fast Gradient Sign Method (FGSM) for adversarial training\n",
    "epsilon = 0.01  # Small magnitude of perturbation\n",
    "\n",
    "# Creating perturbed embeddings by adding small perturbations to the input embeddings\n",
    "### YOUR CODE HERE (perturb the inputs by epsilon in the direction of the input_gradients)\n",
    "# perturbed_embeddings = \n",
    "\n",
    "# Forward pass with perturbed embeddings to evaluate model performance on adversarial example\n",
    "logits, _ = model(src, tgt_input, src_padding_mask, tgt_mask, perturbed_embeddings)\n",
    "loss_after = loss_fn(logits, tgt_out)\n",
    "print(f'loss after perturbation: {loss_after: 0.6f}')\n",
    "\n",
    "percent_increase = 100* (loss_after - loss) / loss\n",
    "print(f\"percent increase in the loss: {percent_increase:0.2f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Manual crafting using synonyms\n",
    "\n",
    "Manual crafting of inputs to test transformer models can be a time-consuming task. However, with the right techniques, we can automate parts of this process for efficiency. Here's a streamlined approach to get you started:\n",
    "\n",
    "##### Step 1: Selecting Target Words\n",
    "- Begin by identifying key words in your sentences that are likely to influence the model's output significantly.\n",
    "\n",
    "##### Step 2: Automated Synonym Replacement\n",
    "- Use natural language processing tools (like NLTK) to find synonyms for these selected words.\n",
    "\n",
    "##### Step 3: Sentence Reconstruction\n",
    "- Integrate these synonyms back into your sentences, ensuring they fit grammatically and contextually.\n",
    "\n",
    "##### Step 4: Testing and Iteration\n",
    "- Feed these modified sentences into your model to observe any changes in output.\n",
    "- Iterate this process, refining your word selection and replacement strategy for optimal results.\n",
    "\n",
    "By following these steps, you can make the process of crafting adversarial or test inputs for transformer models more efficient and less labor-intensive.\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original: The quick brown fox jumps over the lazy dog.\n",
      "Adversarial: The immediate brownuniversity fox jumps over the lazy dog .\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package wordnet to\n",
      "[nltk_data]     /home/mila/g/guptapra/nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n",
      "[nltk_data] Downloading package averaged_perceptron_tagger to\n",
      "[nltk_data]     /home/mila/g/guptapra/nltk_data...\n",
      "[nltk_data]   Package averaged_perceptron_tagger is already up-to-\n",
      "[nltk_data]       date!\n"
     ]
    }
   ],
   "source": [
    "import nltk\n",
    "from nltk.corpus import wordnet\n",
    "nltk.download('wordnet')\n",
    "nltk.download('averaged_perceptron_tagger')\n",
    "\n",
    "def get_synonyms(word):\n",
    "    synonyms = set()\n",
    "    for syn in wordnet.synsets(word):\n",
    "        for lemma in syn.lemmas():\n",
    "            synonym = lemma.name().replace(\"_\", \" \").replace(\"-\", \" \").lower()\n",
    "            synonym = \"\".join([char for char in synonym if char.isalpha()])\n",
    "            synonyms.add(synonym)\n",
    "    if word in synonyms:\n",
    "        synonyms.remove(word)\n",
    "    return list(synonyms)\n",
    "\n",
    "def synonym_replacement(sentence, num_replacements=1):\n",
    "    words = nltk.word_tokenize(sentence)\n",
    "    new_words = words.copy()\n",
    "    replaced = 0\n",
    "    for word in words:\n",
    "        synonyms = get_synonyms(word)\n",
    "        if synonyms:\n",
    "            synonym = random.choice(synonyms)\n",
    "            new_words = [synonym if w == word else w for w in new_words]\n",
    "            replaced += 1\n",
    "            if replaced >= num_replacements:\n",
    "                break\n",
    "    return ' '.join(new_words)\n",
    "\n",
    "# Example\n",
    "original_sentence = \"The quick brown fox jumps over the lazy dog.\"\n",
    "adversarial_sentence = synonym_replacement(original_sentence, num_replacements=2)\n",
    "print(\"Original:\", original_sentence)\n",
    "print(\"Adversarial:\", adversarial_sentence)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualization of attention \n",
    "\n",
    "Visualizing attention in a Transformer model can provide significant insights into how the model processes and interprets input data. However, deciding which layer(s) to visualize depends on what you want to understand or demonstrate. The Transformer architecture, as proposed by Vaswani et al., consists of multiple layers of encoders and decoders, each with its multi-head attention mechanism. Here are some considerations for choosing which layer's attention to visualize:\n",
    "\n",
    "### Encoder Layers:\n",
    "\n",
    "1. **Early Encoder Layers**: These layers often capture more about the syntax and structure of the input text. Visualizing attention here can show how the model is focusing on different parts of the input sentence to understand its grammatical structure.\n",
    "\n",
    "2. **Later Encoder Layers**: These tend to capture more abstract and contextual information. Visualizing attention in these layers can reveal how different words or phrases in the input sentence are related in terms of meaning or context.\n",
    "\n",
    "### Decoder Layers:\n",
    "\n",
    "1. **Early Decoder Layers**: In the decoder, early layers might focus more on aligning with the relevant parts of the input sequence (in the case of tasks like translation). This is where the cross-attention mechanism plays a key role, and visualizing it can show how each word in the output sequence is being aligned with words in the input sequence.\n",
    "\n",
    "2. **Later Decoder Layers**: These layers are more about refining the output and capturing the broader context of the entire sequence. Attention visualizations here can show how the model consolidates information from the input and the partially decoded sequence to produce the final output.\n",
    "\n",
    "### Cross-Attention vs. Self-Attention:\n",
    "\n",
    "- **Cross-Attention (Decoder)**: Specifically in translation tasks, cross-attention layers in the decoder show how each word in the target language is being aligned to words in the source language. This can be crucial for understanding translation decisions made by the model.\n",
    "\n",
    "- **Self-Attention (Both Encoder and Decoder)**: Self-attention in both the encoder and decoder shows dependencies and relationships within the same sequence (source in encoder, target in decoder). \n",
    "\n",
    "### Practical Approach:\n",
    "\n",
    "A practical approach to attention visualization in Transformers would be to start with one layer (e.g., the first or last layer of the encoder), then compare it with other layers to see how attention patterns evolve. This can provide a comprehensive view of how the model's focus shifts from capturing basic syntactic relationships to understanding complex semantic contexts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Following keys are in attn_scores: dict_keys(['encoder_attn', 'decoder_ss', 'decoder_cs'])\n"
     ]
    }
   ],
   "source": [
    "src, tgt = get_sample_batch('train', 1)\n",
    "\n",
    "# Copy-pasted from `process` \n",
    "src = src.to(device)\n",
    "tgt = tgt.to(device)\n",
    "\n",
    "# ignore last token in the tgt because at the time of predicting the last token, available inputs will be everything before that token\n",
    "tgt_input = tgt[:, :-1]\n",
    "\n",
    "# create masks\n",
    "src_padding_mask, tgt_mask = create_masks(src, tgt_input)\n",
    "src_padding_mask = src_padding_mask.to(device)\n",
    "tgt_mask = tgt_mask.to(device)\n",
    "\n",
    "logits, attn_scores = model(src, tgt_input, src_padding_mask, tgt_mask)\n",
    "\n",
    "print(f\"Following keys are in attn_scores: {attn_scores.keys()}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "sentence = get_readable_tokens(src[0, :].cpu(), SRC_LANGUAGE)\n",
    "predicted_sentence = get_readable_tokens(tgt[0, :].cpu(), TGT_LANGUAGE)\n",
    "\n",
    "## Parameters to define specific attention layer\n",
    "batch_idx = 0 # which sentence in the batch\n",
    "layer = 2 # which layer\n",
    "head_idx = 0 # which head\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualizing cross-attention"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Scores shape: (12, 15)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(10,5), dpi=100)\n",
    "\n",
    "scores = attn_scores['decoder_cs'][layer].cpu().detach().numpy()\n",
    "print(f\"Attention scores shape: {scores[batch_idx, head_idx].shape}\")\n",
    "\n",
    "sns.heatmap(scores[batch_idx, head_idx], annot=True,\n",
    "            xticklabels=sentence, yticklabels=predicted_sentence, ax=ax, fmt=\".2f\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualizing self-attention"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Scores shape: (12, 12)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(10,5), dpi=100)\n",
    "\n",
    "scores = attn_scores['decoder_ss'][layer].cpu().detach().numpy()\n",
    "\n",
    "print(f\"Scores shape: {scores[batch_idx, head_idx].shape}\")\n",
    "sns.heatmap(scores[batch_idx, head_idx], annot=True,\n",
    "            xticklabels=predicted_sentence, yticklabels=predicted_sentence, ax=ax, fmt=\".2f\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementing Low-Rank Adaptation (LoRA) for LLMs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section, we'll implement the Low-Rank Adaptation (LoRA) for Large Language Models (LLMs), as proposed in Hu et al. 2021. LoRA focuses on refining pre-trained transformer models without altering the original model weights.\n",
    "\n",
    "\n",
    "\n",
    "#### Overview of LoRA\n",
    "We introduce additional, trainable parameters to the model, which are significantly fewer in number than the original weights. These new parameters are adjusted according to a loss function, using an additional dataset.\n",
    "\n",
    "Here is the description from the paper \n",
    "\n",
    "<img src=\"img/lora_def.png\">\n",
    "\n",
    "#### Steps for Implementation\n",
    "1. **Freezing Pre-trained Parameters**: Ensure all original model parameters are fixed and not updated during training.\n",
    "\n",
    "2. **Calculating Loss with LoRA Weights**:\n",
    "   - Compute the loss function on a sample batch.\n",
    "   - Backpropagate this loss specifically to train the LoRA weights.\n",
    "\n",
    "3. **Iterative Training**:\n",
    "   - Continuously adjust the LoRA weights based on the loss calculation.\n",
    "   - Aim for the model to adapt to new data while retaining its original capabilities.\n",
    "\n",
    "\n",
    "This approach allows for a more efficient and targeted adaptation of LLMs, enhancing their performance on specific tasks or datasets without the need for extensive retraining.\n",
    "\n",
    "\n",
    "[[Hu et al. 2021](!https://arxiv.org/abs/2106.09685)] LoRA: Low-Rank Adaptation of Large Language Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Approach\n",
    "Instead of fine-tuning the entire model, we'll apply LoRA only to the encoder's Q and V. Thus, We'll focus on a targeted adaptation of our transformer model's encoders, specifically modifying the Query (Q) and Value (V) matrices, as suggested by Hu et al. in their paper."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_num_trainable_params(model):\n",
    "    return sum(p.numel() for p in model.parameters() if p.requires_grad)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class LoRA(nn.Module):\n",
    "    def __init__(self, base_module, rank, d_in, d_out):\n",
    "        super(LoRA, self).__init__()\n",
    "        \n",
    "        self.base_module = base_module\n",
    "        \n",
    "        ### YOUR CODE HERE\n",
    "        # self.A = nn.Parameter(torch.randn(?, ?))\n",
    "        # self.B = nn.Parameter(torch.randn(?, ?))        \n",
    "    \n",
    "    def forward(self, x):\n",
    "        out = self.base_module(x)\n",
    "        ### YOUR CODE HERE\n",
    "        # lora_update = \n",
    "        return out + lora_update"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's **freeze all the parameters** so that they are fixed to their pretrained values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of trainable parameters: 6090070\n",
      "encoders.0.mha.Q.weight\n",
      "encoders.0.mha.Q.bias\n",
      "encoders.0.mha.V.weight\n",
      "encoders.0.mha.V.bias\n",
      "encoders.1.mha.Q.weight\n",
      "encoders.1.mha.Q.bias\n",
      "encoders.1.mha.V.weight\n",
      "encoders.1.mha.V.bias\n",
      "encoders.2.mha.Q.weight\n",
      "encoders.2.mha.Q.bias\n",
      "encoders.2.mha.V.weight\n",
      "encoders.2.mha.V.bias\n",
      "Number of trainable parameters: 0\n"
     ]
    }
   ],
   "source": [
    "print(f'Number of trainable parameters: {get_num_trainable_params(model)}')\n",
    "for name, params in model.named_parameters():\n",
    "    params.requires_grad = False\n",
    "    \n",
    "    if 'mha.Q' in name or 'mha.V' in name:\n",
    "        print(name)\n",
    "\n",
    "print(f'Number of trainable parameters: {get_num_trainable_params(model)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's make encoders retrainable by subclassing them with LoRA. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of trainable parameters: 6144\n"
     ]
    }
   ],
   "source": [
    "for encoder in model.encoders:\n",
    "    d_in = encoder.mha.Q.in_features\n",
    "    d_out = encoder.mha.Q.out_features\n",
    "    \n",
    "    ### YOUR CODE HERE \n",
    "    # encoder.mha.Q = \n",
    "    \n",
    "    d_in = encoder.mha.V.in_features\n",
    "    d_out = encoder.mha.V.out_features  \n",
    "    \n",
    "    ### YOUR CODE HERE\n",
    "    # encoder.mha.V = \n",
    "    \n",
    "print(f'Number of trainable parameters: {get_num_trainable_params(model)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we will compute the loss function on a sample batch and backpropagate the loss to train LoRA weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([10, 18, 128]) torch.Size([10, 18, 128])\n",
      "torch.Size([10, 18, 128]) torch.Size([10, 18, 128])\n",
      "torch.Size([10, 18, 128]) torch.Size([10, 18, 128])\n",
      "torch.Size([10, 18, 128]) torch.Size([10, 18, 128])\n",
      "torch.Size([10, 18, 128]) torch.Size([10, 18, 128])\n",
      "torch.Size([10, 18, 128]) torch.Size([10, 18, 128])\n"
     ]
    }
   ],
   "source": [
    "# finetune these parameters\n",
    "loss = LabelSmoothingLoss(TGT_N_TOKENS, 0.1)\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=0.0001, betas=(0.9, 0.98), eps=1e-9)\n",
    "\n",
    "src, tgt = get_sample_batch('train', 10)\n",
    "\n",
    "# Copy-pasted from `process` \n",
    "src = src.to(device)\n",
    "tgt = tgt.to(device)\n",
    "\n",
    "# ignore last token in the tgt because at the time of predicting the last token, available inputs will be everything before that token\n",
    "tgt_input = tgt[:, :-1]\n",
    "\n",
    "# create masks\n",
    "src_padding_mask, tgt_mask = create_masks(src, tgt_input)\n",
    "src_padding_mask = src_padding_mask.to(device)\n",
    "tgt_mask = tgt_mask.to(device)\n",
    "\n",
    "logits, attn_scores = model(src, tgt_input, src_padding_mask, tgt_mask)\n",
    "\n",
    "loss = loss_fn(logits, tgt_input)\n",
    "\n",
    "optimizer.zero_grad()\n",
    "loss.backward()\n",
    "optimizer.step()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bias checking \n",
    "\n",
    "Crafting inputs to check for bias or unethical outputs in a Transformer model requires a careful and systematic approach. Transformers, like any machine learning model, can exhibit biases that are present in their training data. Testing for these biases involves probing the model with inputs designed to reveal discriminatory, unfair, or otherwise unethical responses. \n",
    "\n",
    "\n",
    "1. **Identify Potential Areas of Bias:** Start by identifying dimensions along which biases might occur. Common areas include gender, race, ethnicity, age, sexual orientation, religion, and political viewpoint. Bias can manifest in various forms, including stereotyping, underrepresentation, or overrepresentation of certain groups.\n",
    "\n",
    "\n",
    "2. **Design Test Cases:** Create a set of inputs that specifically target these potential areas of bias. The test cases should be designed to reveal any preferential treatment, stereotyping, or derogatory outputs related to the identified dimensions. Examples include:\n",
    "    - ***Gender Bias***: Inputs comparing professions or activities with different gender pronouns (e.g., \"He is a nurse\" vs. \"She is a nurse\").\n",
    "    - ***Racial or Ethnic Bias***: Sentences describing people of different races or ethnicities in various contexts, checking for differences in sentiment or tone.\n",
    "    - ***Religious or Political Bias***: Statements about different religious or political groups to see if the model responds more favorably or unfavorably to some.\n",
    "\n",
    "\n",
    "3. **Neutral and Comparative Inputs:** For each test case, it's helpful to have a neutral input for comparison. This neutral input should be as similar as possible to the test case but without the specific bias-inducing element.\n",
    "\n",
    "\n",
    "4. **Evaluate the Outputs:** Run these test cases through your Transformer model and evaluate the outputs. Look for:\n",
    "    - ***Differences in Sentiment or Tone***: Is the language noticeably more positive or negative for certain groups?\n",
    "    - ***Use of Stereotypes***: Does the model rely on stereotypes or generalizations?\n",
    "    - ***Differential Treatment***: Are similar actions or qualities described differently based on the group being referred to?\n",
    "\n",
    "\n",
    "5. **Quantitative Analysis:** Where possible, perform a quantitative analysis. For instance, use sentiment analysis tools to get objective measures of the tone of the outputs, or employ statistical tests to see if differences in treatment are significant.\n",
    "\n",
    "6. **Iterative Testing:** Bias testing should be iterative. Based on findings, refine your test cases and probe deeper into areas where biases were uncovered.\n",
    "\n",
    "### Example Test Case for Gender Bias:\n",
    "\n",
    "Let's consider an example for testing gender bias:\n",
    "\n",
    "- **Test Input**: \"The doctor asked the nurse for assistance in the operation.\"\n",
    "- **Neutral Comparative Input**: \"The doctor asked the colleague for assistance in the operation.\"\n",
    "- **Biased Variant**: Switch gender pronouns or names traditionally associated with different genders, e.g., \"The doctor asked him for assistance\" vs. \"The doctor asked her for assistance.\"\n",
    "\n",
    "**Ethical Considerations:** While testing for bias, it's crucial to handle sensitive topics with care and respect. The goal is to make AI more fair and equitable, not to perpetuate stereotypes or harm.\n",
    "\n",
    "Bias testing in Transformer models is an evolving area. Continuous monitoring and testing, coupled with a commitment to ethical AI practices, are necessary to mitigate biases. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "gender_neutral_pairs = [\n",
    "    (\"Die Krankenschwester verabreichte Medizin.\", \"The nurse administered medicine.\"),\n",
    "    (\"Der Ingenieur reparierte das Auto.\", \"The engineer fixed the car.\"),\n",
    "    (\"Ein Koch bereitete das Essen zu.\", \"A chef cooked the meal.\"),\n",
    "    (\"Der Lehrer erklärte die Lektion.\", \"The teacher explained the lesson.\"),\n",
    "    (\"Der Pilot flog das Flugzeug.\", \"The pilot flew the plane.\"),\n",
    "    (\"Die Sekretärin organisierte die Akten.\", \"The secretary organized the files.\"),\n",
    "    (\"Der Programmierer schrieb den Code.\", \"The programmer wrote the code.\"),\n",
    "    (\"Der Wissenschaftler veröffentlichte eine Abhandlung.\", \"The scientist published a paper.\"),\n",
    "    (\"Der Fahrer gewann das Rennen.\", \"The driver won the race.\"),\n",
    "    (\"Der Manager leitete das Treffen.\", \"The manager led the meeting.\")\n",
    "]\n",
    "\n",
    "# If the model is biased, it will produce a translation which is biased towards a specific group. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "Translated sentence:  The nurse is receiving fencing .    \n",
      "Expected sentence: The nurse administered medicine.\n",
      "\n",
      "\n",
      "Translated sentence:  The professional athlete is kicking the car .  \n",
      "Expected sentence: The engineer fixed the car.\n",
      "\n",
      "\n",
      "Translated sentence:  A chef is preparing food .    \n",
      "Expected sentence: A chef cooked the meal.\n",
      "\n",
      "\n",
      "Translated sentence:  The teacher is scooping up the backstroke .  \n",
      "Expected sentence: The teacher explained the lesson.\n",
      "\n",
      "\n",
      "Translated sentence:  The pilot is sweeping the plane .   \n",
      "Expected sentence: The pilot flew the plane.\n",
      "\n",
      "\n",
      "Translated sentence:  The senior citizens is scooping up the sprinkler . \n",
      "Expected sentence: The secretary organized the files.\n",
      "\n",
      "\n",
      "Translated sentence:  The professional athlete is crouched in the marble . \n",
      "Expected sentence: The programmer wrote the code.\n",
      "\n",
      "\n",
      "Translated sentence:  The scientist is wearing a professional uniform is doing a\n",
      "Expected sentence: The scientist published a paper.\n",
      "\n",
      "\n",
      "Translated sentence:  The driver is racing in the race .  \n",
      "Expected sentence: The driver won the race.\n",
      "\n",
      "\n",
      "Translated sentence:  The professional team is tackling the rest of the wedding\n",
      "Expected sentence: The manager led the meeting.\n"
     ]
    }
   ],
   "source": [
    "src, tgt = collate_fn(gender_neutral_pairs)\n",
    "output = greedy_decoding(model, src, 10)\n",
    "output_sentence = [get_readable_text(get_readable_tokens(output[i, :].cpu(), TGT_LANGUAGE)) for i in range(src.shape[0])]\n",
    "\n",
    "for i in range(src.shape[0]):\n",
    "    print(\"\\n\")\n",
    "    print(f'Translated sentence: {output_sentence[i]}')\n",
    "    print(f'Expected sentence: {gender_neutral_pairs[i][1]}')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}