Chi Su

Nicolae Darii, Ranjan Sharma, Vladislav Akhmatov et al.

The modern power system, increasingly composed of Inverter-Based Resources (IBR) from multiple manufacturers, requires new study and design techniques that balance accuracy with the need to protect the Intellectual Property (IP) of various stakeholders. One possible method to support detailed electromagnetic transient (EMT) simulations is to convert the original equipment manufacturers (OEM) models into shareable black-box versions using dynamic link libraries (DLLs). This technique prevents IP violations while potentially maintaining simulation accuracy by embedding the original components within the shareable DLL. Thereby, this work aims explicitly to enhance simulation fidelity by translating full-switching models of offshore wind turbines (OWTs). In this context, the paper offers valuable recommendations, including how to convert interpolation-based elements, preserve simulation speed, recognize limitations, and outline future improvements

Luyu Qiu, Jianing Li, Chi Su et al.

Transformer-based large language models have achieved remarkable performance across various natural language processing tasks. However, they often struggle with seemingly easy tasks like arithmetic despite their vast capabilities. This stark disparity raise human's concerns about their safe and ethical use, hinder their widespread adoption.In this paper, we focus on a typical arithmetic task, integer multiplication, to explore and explain the imperfection of transformers in this domain. We provide comprehensive analysis of a vanilla transformer trained to perform n-digit integer multiplication. Our observations indicate that the model decomposes multiplication task into multiple parallel subtasks, sequentially optimizing each subtask for each digit to complete the final multiplication. Based on observation and analysis, we infer the reasons of transformers deficiencies in multiplication tasks lies in their difficulty in calculating successive carryovers and caching intermediate results, and confirmed this inference through experiments. Guided by these findings, we propose improvements to enhance transformers performance on multiplication tasks. These enhancements are validated through rigorous testing and mathematical modeling, not only enhance transformer's interpretability, but also improve its performance, e.g., we achieve over 99.9% accuracy on 5-digit integer multiplication with a tiny transformer, outperform LLMs GPT-4. Our method contributes to the broader fields of model understanding and interpretability, paving the way for analyzing more complex tasks and Transformer models. This work underscores the importance of explainable AI, helping to build trust in large language models and promoting their adoption in critical applications.