تعتبر Adgedding Word ضرورية لنماذج الشبكة العصبية لمختلف مهام معالجة اللغة الطبيعية. نظرا لأن كلمة تضمينها عادة ما يكون لها حجم كبير، من أجل نشر نموذج شبكة عصبي وجوده على أجهزة Edge، يجب ضغطه بشكل فعال. كانت هناك دراسة لاقتراح طريقة تقريبية منخفضة رتبة بلوك من أجل تضمين كلمة، تسمى GroupReduce. حتى لو كان هيكلهم فعالا، فإن الخصائص وراء مفهوم برنامج تضمين الكلمة الحكيمة غير الحكيمة لم يتم استكشافه بما فيه الكفاية. بدافع من هذا، نحن نحسن Grouppreduce من حيث ترجيح الكلمة والهيت. بالنسبة للتوزيع النصي، نقترح طريقة بسيطة ولكنها فعالة مستوحاة من مصطلح طريقة تردد المستندات العكسية في التردد وطريقة تامة بناء عليهم، نبني كلمة تمييزية تضمين خوارزمية ضغط. في التجارب، نوضح أن الخوارزمية المقترحة تجد بشكل أكثر فعالية أوزان الكلمات أكثر من المنافسين في معظم الحالات. بالإضافة إلى ذلك، نوضح أن الخوارزمية المقترحة يمكن أن تتصرف مثل إطار من خلال التعاون الناجح مع الكمي.
Word embedding is essential for neural network models for various natural language processing tasks. Since the word embedding usually has a considerable size, in order to deploy a neural network model having it on edge devices, it should be effectively compressed. There was a study for proposing a block-wise low-rank approximation method for word embedding, called GroupReduce. Even if their structure is effective, the properties behind the concept of the block-wise word embedding compression were not sufficiently explored. Motivated by this, we improve GroupReduce in terms of word weighting and structuring. For word weighting, we propose a simple yet effective method inspired by the term frequency-inverse document frequency method and a novel differentiable method. Based on them, we construct a discriminative word embedding compression algorithm. In the experiments, we demonstrate that the proposed algorithm more effectively finds word weights than competitors in most cases. In addition, we show that the proposed algorithm can act like a framework through successful cooperation with quantization.
References used
https://aclanthology.org/
Neural sequence-to-sequence (Seq2Seq) models and BERT have achieved substantial improvements in abstractive document summarization (ADS) without and with pre-training, respectively. However, they sometimes repeatedly attend to unimportant source phra
The adoption of Transformer-based models in natural language processing (NLP) has led to great success using a massive number of parameters. However, due to deployment constraints in edge devices, there has been a rising interest in the compression o
Cross-Lingual Word Embeddings (CLWEs) encode words from two or more languages in a shared high-dimensional space in which vectors representing words with similar meaning (regardless of language) are closely located. Existing methods for building high
Word embeddings are powerful representations that form the foundation of many natural language processing architectures, both in English and in other languages. To gain further insight into word embeddings, we explore their stability (e.g., overlap b
Sarcasm detection is of great importance in understanding people's true sentiments and opinions. Many online feedbacks, reviews, social media comments, etc. are sarcastic. Several researches have already been done in this field, but most researchers