Is Traditional Machine Learning Useless?

Given the Boom of Deep Learning Approaches

Mario Graff

INFOTEC

INFOTEC

Aguascalientes, México

INFOTEC

Centro Público de Investigación del Gobierno Federal, que contribuye a la Transformación Digital de México, a través de la investigación, la innovación, la formación académica y el desarrollo de productos y servicios TIC. Sus alcances abarcan al sector público y privado, habilitando caminos que conduzcan hacia un México moderno y de inclusión digital.

INGEOTEC

GitHub: https://github.com/INGEOTEC
WebPage: https://ingeotec.github.io/

Introduction

Definitions

Artificial Intelligence (AI)

Theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Machine Learning

Machine learning (ML) is a subfield of artificial intelligence that focuses on the development and implementation of algorithms capable of learning from data without being explicitly programmed.

Natural Language Processing (NLP)

NLP is a branch of artificial intelligence (AI) that uses machine learning and other technologies to enable computers to understand, process, and manipulate human language.

AI Perception

How we look at it

How I see it

Machine Learning

Linear Classifier

Decision Tree

Problem

Classifier

Natural Language Processing

Text Classification

Definition

The aim is the classification of documents into a fixed number of predefined categories.

Polarity

El día de mañana no podré ir con ustedes a la librería

Negative

Training set

https://ingeotec.github.io/Delitos

texto etiqueta
0 Jorge Lanata sobre "la detención de Boudou", ... N
1 #31Jul #Sucesos #NiUnaMenos \n52 mujeres fuero... P
2 @anhetch @SilvieChavez En este caso el débil e... N
3 Camión atropella a motochorros en gasolinera d... P
4 Qué recuerdos cuando antes de irme al colegio ... N
5 #VIOLENCIA \nDe la violencia que se registra e... P
6 @osa409 @memohiervas Según vídeo que eh visto,... N
7 Reportan una persona ejecutada con disparo de ... P

Linear Classifier

Text Representation

Bag of Words

Text Representation (2)

Associate token \(t\)

\[ \mathbf{v_t} \in \mathbb R^d \]

Bag of words

\[ \mathbf x = \frac{\sum_t \mathbf{v_t}}{\lVert \sum_t \mathbf{v_t} \rVert} \]

Orthogonal

\[ \forall_{i \neq j} \mathbf{v_i} \cdot \mathbf{v_j} = 0 \]

Consequences

  • No similarity between tokens
  • High dimension

Text Representation (3)

Document / TFIDF

Segmentación

from encexp.text_repr import SeqTM
seq = SeqTM(lang='es')
seq.tokenize('Es un placer estar con ustedes')
['es~un', 'placer', 'estar~con', 'ustedes']

Representación

seq['Es un placer estar con ustedes']
[(4957, 0.36521129661206075),
 (8374, 0.5522348027783196),
 (4785, 0.5932117431782478),
 (15783, 0.4579926409700842)]

Self-supervised Learning

Select a token

Supervised learning

Self-supervised Learning (2)

Classification

CBOW

  • \(\mathbf h = s(A \mathbf x)\)
  • \(\mathbf y = s(A_2 \mathbf h)\)
  • \(\mathbf x \in \mathbb{R}^d\), \(A \in \mathbb{R}^{(300, d)}\),
  • \(\mathbf h \in \mathbb{R}^{300}\), and \(A_2 \in \mathbb{R}^{(d, 300)}\)

Dense representation

  • Columns of \(A\)

Self-supervised Learning (3)

Classification

Linear Classifier

  • \(y_t = \textsf{sign}(\mathbf w_t \mathbf x)\)
  • \(\mathbf x \in \mathbb{R}^d\), \(\mathbf w_t \in \mathbb{R}^d\)

Dense representation

  • \(\mathbf w_t\)
  • e.g., \(d = 2^{13}\), and \(t=1,\ldots,2^{13}\)
  • \(W \approx\) 67M

Self-supervised Learning (4)

Classification

Outside NLP

  • Missing values
  • High dimension / rare events

Contextual Representation

Attention is All you Need

(Vaswani et al. 2017)

BERT

(Devlin et al. 2019)

Attention

Equation

\[ \textsf{att}(Q, K, V) = \textsf{softmax}(\frac{QK^\intercal}{\sqrt{d_k}}) V \]

Parts

  • \(Q^{(\ell, d_a)} = H^{(\ell, d_h)}Q^{(d_h, d_a)}\)
  • \(K^{(\ell, d_a)} = H^{(\ell, d_h)}K^{(d_h, d_a)}\)
  • \(V^{(\ell, d_a)} = H^{(\ell, d_h)}V^{(d_h, d_a)}\)

Analysis

  • \(W^{(\ell, \ell)} = \textsf{softmax}(\frac{QK^\intercal}{\sqrt{d_k}})\)
  • \(\textsf{att}(Q, K, V)^{(\ell, d_a)} = W^{(\ell, \ell)} V^{(\ell, d_a)}\)
  • \(H^{(\ell, d_h)} \rightarrow \textsf{att}(Q, K, V)^{(\ell, d_a)}\)

Performance

Author Profiling

English / Italian

Spanish

Conclusions

Definiciones

  • Artificial Intelligence
  • Machine Learning
  • Text Classification
  • Bag of Words
  • Sparse Representation
  • Dense BoW

Referencias

Bormuth, John R. 1968. “Cloze Test Readability: Criterion Reference Scores.” Journal of Educational Measurement 5 (3): 189–96.
Devlin, Jacob, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. “BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding.” In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), edited by Jill Burstein, Christy Doran, and Thamar Solorio, 4171–86. Minneapolis, Minnesota: Association for Computational Linguistics. https://doi.org/10.18653/v1/N19-1423.
Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. “Attention Is All You Need.” In Advances in Neural Information Processing Systems, edited by I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett. Vol. 30. Curran Associates, Inc. https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf.

¡Gracias!