Hongyi Wu

Nuowei Liu, Xinhao Chen, Hongyi Wu et al.

Existing rhetorical understanding and generation datasets or corpora primarily focus on single coarse-grained categories or fine-grained categories, neglecting the common interrelations between different rhetorical devices by treating them as independent sub-tasks. In this paper, we propose the Chinese Essay Rhetoric Dataset (CERD), consisting of 4 commonly used coarse-grained categories including metaphor, personification, hyperbole and parallelism and 23 fine-grained categories across both form and content levels. CERD is a manually annotated and comprehensive Chinese rhetoric dataset with five interrelated sub-tasks. Unlike previous work, our dataset aids in understanding various rhetorical devices, recognizing corresponding rhetorical components, and generating rhetorical sentences under given conditions, thereby improving the author's writing proficiency and language usage skills. Extensive experiments are conducted to demonstrate the interrelations between multiple tasks in CERD, as well as to establish a benchmark for future research on rhetoric. The experimental results indicate that Large Language Models achieve the best performance across most tasks, and jointly fine-tuning with multiple tasks further enhances performance.

Xiangrui Xu, Qiao Zhang, Rui Ning et al.

The prevalent use of Transformer-like models, exemplified by ChatGPT in modern language processing applications, underscores the critical need for enabling private inference essential for many cloud-based services reliant on such models. However, current privacy-preserving frameworks impose significant communication burden, especially for non-linear computation in Transformer model. In this paper, we introduce a novel plug-in method Comet to effectively reduce the communication cost without compromising the inference performance. We second introduce an efficient approximation method to eliminate the heavy communication in finding good initial approximation. We evaluate our Comet on Bert and RoBERTa models with GLUE benchmark datasets, showing up to 3.9$\times$ less communication and 3.5$\times$ speedups while keep competitive model performance compared to the prior art.

T. Luo, S. S. Kanhere, H-P. Tan et al.

Incentive mechanisms for crowdsourcing have been extensively studied under the framework of all-pay auctions. Along a distinct line, this paper proposes to use Tullock contests as an alternative tool to design incentive mechanisms for crowdsourcing. We are inspired by the conduciveness of Tullock contests to attracting user entry (yet not necessarily a higher revenue) in other domains. In this paper, we explore a new dimension in optimal Tullock contest design, by superseding the contest prize---which is fixed in conventional Tullock contests---with a prize function that is dependent on the (unknown) winner's contribution, in order to maximize the crowdsourcer's utility. We show that this approach leads to attractive practical advantages: (a) it is well-suited for rapid prototyping in fully distributed web agents and smartphone apps; (b) it overcomes the disincentive to participate caused by players' antagonism to an increasing number of rivals. Furthermore, we optimize conventional, fixed-prize Tullock contests to construct the most superior benchmark to compare against our mechanism. Through extensive evaluations, we show that our mechanism significantly outperforms the optimal benchmark, by over three folds on the crowdsourcer's utility cum profit and up to nine folds on the players' social welfare.