Skip to main content

Understanding Distance Metrics in Machine Learning with PyTorch Examples

Distance metrics play a crucial role in machine learning, especially in tasks like clustering, classification, and recommendation systems. In this blog, we will explore popular distance metrics including Cosine, Euclidean, Mahalanobis, Hellinger, Jaccard, Manhattan, Correlation, Dice, and Hamming distances. We will also provide PyTorch implementations for each metric.

1. Cosine Distance

Measures the cosine of the angle between two non-zero vectors. Often used in text similarity and document clustering.

import torch
x = torch.tensor([1.0, 2.0, 3.0])
y = torch.tensor([4.0, 5.0, 6.0])
cosine_distance = 1 - torch.nn.functional.cosine_similarity(x.unsqueeze(0), y.unsqueeze(0))

2. Euclidean Distance

Represents the straight-line distance between two points in Euclidean space.

euclidean_distance = torch.dist(x, y, p=2)

3. Mahalanobis Distance

Accounts for the correlation between variables and scales distances accordingly. Useful in anomaly detection.

cov = torch.cov(torch.stack([x, y]).T)
cov_inv = torch.linalg.inv(cov)
diff = (x - y).unsqueeze(0)
mahalanobis_distance = torch.sqrt(diff @ cov_inv @ diff.T)

4. Hellinger Distance

Measures the similarity between two probability distributions.

px = torch.sqrt(x / x.sum())
py = torch.sqrt(y / y.sum())
hellinger_distance = torch.norm(px - py) / torch.sqrt(torch.tensor(2.0))

5. Jaccard Distance

Used for comparing similarity and diversity of sample sets. Defined as 1 - (intersection / union).

x_set = torch.tensor([1, 1, 0, 0])
y_set = torch.tensor([1, 0, 1, 0])
intersection = torch.sum((x_set & y_set).float())
union = torch.sum((x_set | y_set).float())
jaccard_distance = 1 - intersection / union

6. Manhattan Distance

Also known as L1 distance. The sum of absolute differences between corresponding elements.

manhattan_distance = torch.sum(torch.abs(x - y))

7. Correlation Distance

Measures dissimilarity between variables by 1 minus the Pearson correlation coefficient.

correlation_distance = 1 - torch.corrcoef(torch.stack([x, y]))[0, 1]

8. Dice Distance

Mainly used in comparing similarity between two sets. Defined as 1 - (2 * |A ∩ B| / (|A| + |B|)).

intersection = torch.sum((x_set & y_set).float())
dice_distance = 1 - (2 * intersection) / (x_set.sum() + y_set.sum())

9. Hamming Distance

Measures the number of positions at which corresponding elements differ.

hamming_distance = torch.sum(x_set != y_set).float() / x_set.numel()

References

  • https://pytorch.org/docs/stable/index.html
  • https://en.wikipedia.org/wiki/Distance_metric
  • https://scikit-learn.org/stable/modules/generated/sklearn.metrics.pairwise_distances.html
  • https://en.wikipedia.org/wiki/Cosine_similarity

Comments

Post a Comment

Popular

Understanding SentencePiece: A Language-Independent Tokenizer for AI Engineers

In the realm of Natural Language Processing (NLP), tokenization plays a pivotal role in preparing text data for machine learning models. Traditional tokenization methods often rely on language-specific rules and pre-tokenized inputs, which can be limiting when dealing with diverse languages and scripts. Enter SentencePiece—a language-independent tokenizer and detokenizer designed to address these challenges and streamline the preprocessing pipeline for neural text processing systems. What is SentencePiece? SentencePiece is an open-source tokenizer and detokenizer developed by Google, tailored for neural-based text processing tasks such as Neural Machine Translation (NMT). Unlike conventional tokenizers that depend on whitespace and language-specific rules, SentencePiece treats the input text as a raw byte sequence, enabling it to process languages without explicit word boundaries, such as Japanese, Chinese, and Korean. This approach allows SentencePiece to train subword models di...
Post a Comment
Read more

Mastering the Byte Pair Encoding (BPE) Tokenizer for NLP and LLMs

Byte Pair Encoding (BPE) is one of the most important and widely adopted subword tokenization algorithms in modern Natural Language Processing (NLP), especially in training Large Language Models (LLMs) like GPT. This guide provides a deep technical dive into how BPE works, compares it with other tokenizers like WordPiece and SentencePiece, and explains its practical implementation with Python code. This article is optimized for AI engineers building real-world models and systems. 1. What is Byte Pair Encoding? BPE was originally introduced as a data compression algorithm by Gage in 1994. It replaces the most frequent pair of bytes in a sequence with a single, unused byte. In 2015, Sennrich et al. adapted BPE for NLP to address the out-of-vocabulary (OOV) problem in neural machine translation. Instead of working with full words, BPE decomposes them into subword units that can be recombined to represent rare or unseen words. 2. Why Tokenization Matters in LLMs Tokenization is th...
Post a Comment
Read more

ZeRO: Deep Memory Optimization for Training Trillion-Parameter Models

In 2020, Microsoft researchers introduced ZeRO (Zero Redundancy Optimizer) via their paper "ZeRO: Memory Optimization Towards Training Trillion Parameter Models" (arXiv:1910.02054). ZeRO is a memory optimization technique that eliminates redundancy in distributed training, enabling efficient scaling to trillion-parameter models. This provides an in-depth technical breakdown of ZeRO's partitioning strategies, memory usage analysis, and integration with DeepSpeed. 1. What is ZeRO? ZeRO eliminates redundant memory copies of model states across GPUs. Instead of replicating parameters, gradients, and optimizer states across each GPU, ZeRO partitions them across all devices. This results in near-linear memory savings as the number of GPUs increases. 2. Limitations of Traditional Data Parallelism In standard data-parallel training, every GPU maintains: Model Parameters $\theta$ Gradients $\nabla \theta$ Optimizer States $O(\theta)$ This causes memory usage ...
Post a Comment
Read more