Searcharxiv⌕ Search

arXiv subjects

Sattaya Singkul

Publications and source records attributed to Sattaya Singkul.

3 recordsLinked to original sources

The Vocabulary Tax: Safely Trimming Unused Tokens from LLMs for On-Device Use

Open large language models (LLMs) and speech models ship vocabularies of up to 262k tokens, and every user downloads the tables those tokens index (the input embedding, an untied output head and, in recent on-device models, per-layer embeddings), whatever languages they need. We call the bytes in rows a deployment never uses the vocabulary tax and measure it, without downloading weights, for 1,244 generative and speech-recognition models on the Hugging Face Hub. For a Thai and English deployment it is a median 4.4% of model bytes below 0.5B parameters and 2.1% above 8B, and 1.93 GB in Gemma 4 E2B: it falls hardest on the small models that edge devices run. Keeping only the rows of the deployment's writing systems is exact. Keeping the tokens seen in the user's own text removes much more, a median 29.0% below 0.5B, but the usual way of doing it breaks BPE tokenizers: it drops the intermediate tokens that merges build on, so new text is split differently 1.9 times as often as the standard check reports. We repair this with merge closure, at a cost of 5.9% more rows, and add a trim certificate: a Good-Turing estimate, from the trimming text alone, of how much new text the trimmed tokenizer will split differently (correlation 0.996 over 300 settings in 22 languages) and of how much more text a target needs. Without test data our method never exceeds a 2% target, while prior keep rules checked on test data exceed it in 9-13 of 44 settings. On six models of 270M to 3B parameters in 11 languages, trimmed models keep their predictions: where the certificate passes, a median 89.7% of rows is removed and bits per character rise by at most 2.0%. The census, the trimming and the certificate are released as an open Python package, and the per-model tax for 41 languages as an open dataset.

cs.CL↗

Deep Residual Local Feature Learning for Speech Emotion Recognition

Speech Emotion Recognition (SER) is becoming a key role in global business today to improve service efficiency, like call center services. Recent SERs were based on a deep learning approach. However, the efficiency of deep learning depends on the number of layers, i.e., the deeper layers, the higher efficiency. On the other hand, the deeper layers are causes of a vanishing gradient problem, a low learning rate, and high time-consuming. Therefore, this paper proposed a redesign of existing local feature learning block (LFLB). The new design is called a deep residual local feature learning block (DeepResLFLB). DeepResLFLB consists of three cascade blocks: LFLB, residual local feature learning block (ResLFLB), and multilayer perceptron (MLP). LFLB is built for learning local correlations along with extracting hierarchical correlations; DeepResLFLB can take advantage of repeatedly learning to explain more detail in deeper layers using residual learning for solving vanishing gradient and reducing overfitting; and MLP is adopted to find the relationship of learning and discover probability for predicted speech emotions and gender types. Based on two available published datasets: EMODB and RAVDESS, the proposed DeepResLFLB can significantly improve performance when evaluated by standard metrics: accuracy, precision, recall, and F1-score.

cs.SD↗

Parsing Thai Social Data: A New Challenge for Thai NLP

Dependency parsing (DP) is a task that analyzes text for syntactic structure and relationship between words. DP is widely used to improve natural language processing (NLP) applications in many languages such as English. Previous works on DP are generally applicable to formally written languages. However, they do not apply to informal languages such as the ones used in social networks. Therefore, DP has to be researched and explored with such social network data. In this paper, we explore and identify a DP model that is suitable for Thai social network data. After that, we will identify the appropriate linguistic unit as an input. The result showed that, the transition based model called, improve Elkared dependency parser outperform the others at UAS of 81.42%.

cs.CL↗