Xiaohan Ma

Large language models (LLMs) show limited capability in processing low-resource historical languages due to insufficient training data and domain-specific linguistic structures. Korean Literary Sinitic (KLS), the principal written medium of the Joseon dynasty, remains particularly under-resourced despite its lexical overlap with modern Korean and shared script with classical Chinese. To enable systematic evaluation in this domain, we introduce KLSBench, a comprehensive benchmark for assessing LLM performance on KLS. KLSBench contains 7871 instances sourced from Joseon dynasty civil service examination archives and parallel corpora of the Four Books, and encompasses five task categories: classification, retrieval, punctuation restoration, natural language inference, and translation. Our evaluation suggests KLSBench could work as an effective diagnostic tool that distinguishes lexical recall from deeper linguistic comprehension in low-resource historical languages. Beyond establishing evaluation baselines, KLSBench provides practical frameworks for deploying LLM-based tools in digital humanities contexts, including automated annotation systems and intelligent search interfaces for classical text repositories.

Sign Language Production (SLP) refers to the task of translating textural forms of spoken language into corresponding sign language expressions. Sign languages convey meaning by means of multiple asynchronous articulators, including manual and non‐manual information channels. Recent deep learning‐based SLP models directly generate the full‐articulatory sign sequence from the text input in an end‐to‐end manner. However, these models largely down weight the importance of subtle differences in the manual articulation due to the effect of regression to the mean. To explore these neglected aspects, an efficient cascade dual‐decoder Transformer (CasDual‐Transformer) for SLP is proposed to learn, successively, two mappings SLP _hand : Text → Hand pose and SLP _sign : Text → Sign pose , utilising an attention‐based alignment module that fuses the hand and sign features from previous time steps to predict more expressive sign pose at the current time step. In addition, to provide more efficacious guidance, a novel spatio‐temporal loss to penalise shape dissimilarity and temporal distortions of produced sequences is introduced. Experimental studies are performed on two benchmark sign language datasets from distinct cultures to verify the performance of the proposed model. Both quantitative and qualitative results show that the authors’ model demonstrates competitive performance compared to state‐of‐the‐art models, and in some cases, achieves considerable improvements over them.

Generative adversarial networks (GANs) have shown significant progress in generating highquality visual samples, however they are still well known both for being unstable to train and for the problem of mode collapse, particularly when trained on data collections containing a diverse set of visual objects. In this paper, we propose an Adaptive k-step Generative Adversarial Network (Ak-GAN), which is designed to mitigate the impact of instability and saturation in the original by dynamically adjusting the ratio of the training steps of both the generator and discriminator. To accomplish this, we track and analyze stable training curves of relatively narrow datasets and use them as the target fitting lines when training more diverse data collections. Furthermore, we conduct experiments on the proposed procedure using several optimization techniques (e.g., supervised guiding from previous stable learning curves with and without momentum) and compare their performance with that of state-of-the-art models on the task of image synthesis from datasets consisting of diverse images. Empirical results demonstrate that Ak-GAN works well in practice and exhibits more stable behavior than regular GANs during training. A quantitative evaluation has been conducted on the Inception Score (IS) and the relative inverse Inception Score (RIS); compared with regular GANs, the former has been improved by 61% and 83%, and the latter by 21% and 60%, on the CelebA and the Anime datasets, respectively.