Real-Time Optimized N-Gram for Mobile Devices

Mani, Sharmila; Gothe, Sourabh Vasant; Ghosh, Sourav; Mishra, Ajay Kumar; Kulshreshtha, Prakhar; Bhargavi, Manju; Kumaran, M. Senthil

doi:10.1109/icosc.2019.8665639

Cited by 7 publications

(6 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We use Op-Ngram [9] as the underlying language model (LM) for suggestions, which has a vocabulary of top-100k words for each language, stored in a Marisa Trie [26], ranked according to their statistical probability in the training corpus. This trie is referred to as VocabTrie.…”

Section: घरmentioning

confidence: 99%

“…where n c denotes the number of actual characters in the test set and n k denotes the number of keystrokes that were needed to produce the test set with a soft keyboard. The NWP metric [9] refers to the probability that given a sequence of the first (k − 1) words from a sentence in test data, the next word in that sentence, w k , will appear among the word prediction candidates. Mathematically,…”

Section: E Evaluationmentioning

confidence: 99%

“…As in Section III, we use Op-Ngram [9] as base LM. The KSR and NWP metrics (Section III-E) are used to compare the improvement due to proposals described in Sections IV-C and IV-D.…”

Section: E Evaluationmentioning

confidence: 99%

“…To the best of our knowledge, no similar solution has been proposed in the literature. We benchmark these approaches using public datasets and Op-Ngram [9] as the base language model. Throughout this work, we use Indic languages and one non-Indic language (Thai) for examples and evaluations; however, the proposals can be directly extended and applied to the entire class of languages with abugida scripts.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

edATLAS: An Efficient Disambiguation Algorithm for Texting in Languages with Abugida Scripts

Ghosh

Gothe

Sanchi

et al. 2021

2021 IEEE 15th International Conference on Semantic Computing (ICSC)

Self Cite

View full text Add to dashboard Cite

Abugida refers to a phonogram writing system where each syllable is represented using a single consonant or typographic ligature, along with a default vowel or optional diacritic(s) to denote other vowels. However, texting in these languages has some unique challenges in spite of the advent of devices with soft keyboard supporting custom key layouts. The number of characters in these languages is large enough to require characters to be spread over multiple views in the layout. Having to switch between views many times to type a single word hinders the natural thought process. This prevents popular usage of native keyboard layouts. On the other hand, supporting romanized scripts (native words transcribed using Latin characters) with language model based suggestions is also set back by the lack of uniform romanization rules.To this end, we propose a disambiguation algorithm and showcase its usefulness in two novel mutually non-exclusive input methods for languages natively using the abugida writing system: (a) disambiguation of ambiguous input for abugida scripts, and (b) disambiguation of word variants in romanized scripts. We benchmark these approaches using public datasets, and show an improvement in typing speed by 19.49%, 25.13%, and 14.89%, in Hindi, Bengali, and Thai, respectively, using Ambiguous Input, owing to the human ease of locating keys combined with the efficiency of our inference method. Our Word Variant Disambiguation (WDA) maps valid variants of romanized words, previously treated as Out-of-Vocab, to a vocabulary of 100k words with high accuracy, leading to an increase in Error Correction F1 score by 10.03% and Next Word Prediction (NWP) by 62.50% on average.

show abstract

Section: घरmentioning

confidence: 99%

Section: E Evaluationmentioning

confidence: 99%

“…As in Section III, we use Op-Ngram [9] as base LM. The KSR and NWP metrics (Section III-E) are used to compare the improvement due to proposals described in Sections IV-C and IV-D.…”

Section: E Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

edATLAS: An Efficient Disambiguation Algorithm for Texting in Languages with Abugida Scripts

Ghosh

Gothe

Sanchi

et al. 2021

2021 IEEE 15th International Conference on Semantic Computing (ICSC)

Self Cite

View full text Add to dashboard Cite

show abstract

“…A language model (LM) calculates the probability of the current word or character with a previous word or character sequence [1]. For the sequence = ( 1 , 2 , … , ), the probability of the LM is denoted as ( ).…”

Section: Introductionmentioning

confidence: 99%

Robustness of Differentiable Neural Computer Using Limited Retention Vector-based Memory Deallocation in Language Model

2021

KSII TIIS

View full text Add to dashboard Cite

Recurrent neural network (RNN) architectures have been used for language modeling (LM) tasks that require learning long-range word or character sequences. However, the RNN architecture is still suffered from unstable gradients on long-range sequences. To address the issue of long-range sequences, an attention mechanism has been used, showing state-of-theart (SOTA) performance in all LM tasks. A differentiable neural computer (DNC) is a deep learning architecture using an attention mechanism. The DNC architecture is a neural network augmented with a content-addressable external memory. However, in the write operation, some information unrelated to the input word remains in memory. Moreover, DNCs have been found to perform poorly with low numbers of weight parameters. Therefore, we propose a robust memory deallocation method using a limited retention vector. The limited retention vector determines whether the network increases or decreases its usage of information in external memory according to a threshold. We experimentally evaluate the robustness of a DNC implementing the proposed approach according to the size of the controller and external memory on the enwik8 LM task. When we decreased the number of weight parameters by 32.47%, the proposed DNC showed a low bits-per-character (BPC) degradation of 4.30%, demonstrating the effectiveness of our approach in language modeling tasks.

show abstract

N-gram models for Text Generation in Hindi Language

Ghude,

Chauhan,

Dahake

et al. 2022

ITM Web Conf.

View full text Add to dashboard Cite

Native language plays a vital role for communication. Hindi is preferred by most Indians and it is the fifth most spoken language in the world. Hence, to make User Experience more effective while interacting with Software Applications, we aim to build a Model using Natural Language Processing which takes a specific word as an input and predicts the subsequent words for completing the sentence.It will act as a tab- complete function in Hindi language. This will also pave a way as a use case for building chatbots, personal emails, applications which are content based like cooking recipes, blogs, etc. in Hindi Language.

show abstract

Real-Time Optimized N-Gram for Mobile Devices

Cited by 7 publications

References 6 publications

edATLAS: An Efficient Disambiguation Algorithm for Texting in Languages with Abugida Scripts

edATLAS: An Efficient Disambiguation Algorithm for Texting in Languages with Abugida Scripts

Robustness of Differentiable Neural Computer Using Limited Retention Vector-based Memory Deallocation in Language Model

N-gram models for Text Generation in Hindi Language

Contact Info

Product

Resources

About