paper-with-me

Papers

Simple Hack for Transformers against Heavy Long-Text Classification on a Time- and Memory-Limited GPU Service

2024-03-19 · Mirza Alim Mutasodirin, Radityo Eko Prasojo, Achmad F. Abka, Hanif Rasyidi

Many NLP researchers rely on free computational services, such as Google Colab, to fine-tune their Transformer models, causing a limitation for hyperparameter optimization (HPO) in long-text classification due to the method having quadratic complexity and needing a bigger resource. In Indonesian, only a few works were found on long-text classification using Transformers. Most only use a small amount of data and do not report any HPO. In this study, using 18k news articles, we investigate which pretrained models are recommended to use based on the output length of the tokenizer. We then compare some hacks to shorten and enrich the sequences, which are the removals of stopwords, punctuation, low-frequency words, and recurring words. To get a fair comparison, we propose and run an efficient and dynamic HPO procedure that can be done gradually on a limited resource and does not require a long-running optimization library. Using the best hack found, we then compare 512, 256, and 128 tokens length. We find that removing stopwords while keeping punctuation and low-frequency words is the best hack. Some of our setups manage to outperform taking 512 first tokens using a smaller 128 or 256 first tokens which manage to represent the same information while requiring less computational resources. The findings could help developers to efficiently pursue optimal performance of the models using limited resources.

📄 PDF Abstract BibTeX arXiv:2403.12563

Code (0)

등록된 구현이 없습니다.

Tasks

ArticlesGPUHyperparameter Optimizationtext-classificationText Classification

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Multi-Head Attention 설명 없음
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Residual Connection 설명 없음

Similar Papers 제목 키워드 기반

HACK: Homomorphic Acceleration via Compression of the Key-Value Cache for Disaggregated LLM Inference

2025-02-05 · Zeyu Zhang, Haiying Shen, Shay Vargaftik, Ran Ben Basat 외

Disaggregated Large Language Model (LLM) inference has gained popularity as it separates the computation-intensive prefill stage from the memory-intensive decode stage, avoiding the prefill-decode interference and improv…

Language ModelingLanguage ModellingLarge Language ModelQuantization

Predicting Organizational Cybersecurity Risk: A Deep Learning Approach

2020-12-26 · Benjamin M. Ampel

Cyberattacks conducted by malicious hackers cause irreparable damage to organizations, governments, and individuals every year. Hackers use exploits found on hacker forums to carry out complex cyberattacks, making explor…

BIG-bench Machine LearningDeep Learning

Quantifying Loss Aversion in Cyber Adversaries via LLM Analysis

2025-08-18 · Soham Hans, Nikolos Gurney, Stacy Marsella, Sofia Hirschmann arxiv

Understanding and quantifying human cognitive biases from empirical data has long posed a formidable challenge, particularly in cybersecurity, where defending against unknown adversaries is paramount. Traditional cyber d…

From Curiosity to Caution: Mitigating Reward Hacking for Best-of-N with Pessimism

2026-04-06 · Zhuohao Yu, Zhiwei Steven Wu, Adam Block arxiv

Inference-time compute scaling has emerged as a powerful paradigm for improving language model performance on a wide range of tasks, but the question of how best to use the additional compute remains open. A popular appr…

Harness-agnostic detection and immunization of reward hacking in self-evolving language models

2026-09-04 · Rongxin Yang, Yang Liu, Shang Luo, Haoxuan Jia 외 arxiv

Self-evolving language models improve by proposing candidate updates and keeping whatever raises a visible score. When that score is an imperfect proxy for the capability one actually wants, sustained selection widens th…