In this article, we explore the capability of speculative fiction to predict future realized technologies. We review a large set of speculative technologies introduced in speculative fiction to determine if the technologies were subsequently realized. Additionally, we explore the time between the speculated introduction and actual realization. Our dataset for analysis is built from the ‘Technovelgy’ database of speculative technologies. A realization assessment methodology is created that includes detailed rubrics to rate and quantify the predictability or realizability of speculative technologies. Three independent raters perform realization assessments for each entry. An inter-rater agreement analysis is carried out to validate the rating method. Based on the dataset of 3095 speculated technologies, 45% are labeled as ‘realized’ by at least one rater. A moderate overall agreement with a Fleiss’ Kappa of 0.57 is reached by all raters. The average time to realization of realized technologies is approximately 45 years with a standard deviation of approximately 34 years. We observe patterns in the realization of speculative technologies and analyze the underlying reasons preventing the technologies from realization. We conclude that speculative fiction predicts future technologies to such a degree that the introduction of speculative technology can be used as an input to designer decision-making.

Large language models (LLMs) have recently been shown to deliver impressive performance in various NLP tasks. To tackle multi-step reasoning tasks, few-shot chain-of-thought (CoT) prompting includes a few manually crafted step-by-step reasoning demonstrations which enable LLMs to explicitly generate reasoning steps and improve their reasoning task accuracy. To eliminate the manual effort, Zeroshot-CoT concatenates the target problem statement with "Let's think step by step" as an input prompt to LLMs. Despite the success of Zero-shot-CoT, it still suffers from three pitfalls: calculation errors, missing-step errors, and semantic misunderstanding errors. To address the missing-step errors, we propose Planand-Solve (PS) Prompting. It consists of two components: first, devising a plan to divide the entire task into smaller subtasks, and then carrying out the subtasks according to the plan. To address the calculation errors and improve the quality of generated reasoning steps, we extend PS prompting with more detailed instructions and derive PS+ prompting. We evaluate our proposed prompting strategy on ten datasets across three reasoning problems. The experimental results over GPT-3 show that our proposed zero-shot prompting consistently outperforms Zero-shot-CoT across all datasets by a large margin, is comparable to or exceeds Zero-shot-Program-of-Thought Prompting, and has comparable performance with 8-shot CoT prompting on the math reasoning problem. The code can be found at https://github.com/AGI-Edgerunners/Plan-and-Solve-Prompting.

Abstract. This paper presents a method of model construction for the power system transient stability assessment based on statistical learning theory integrated with the bagging and the approximate reasoning. Support vector machines operate on the principle of structure risk minimization. This paper takes full advantage of its ability to solve the problem with small sample, nonlinear and high dimension. Hence better generalization ability is guaranteed. The multi-class identification for power system transient stability assessment is solved by the data set reconstruction. The assessment model uses the data set regulation, bagging and approximate reasoning to improve the training speed, the accuracy and stability of the estimation result. The IEEE 39-Bus test system is employed to demonstrate the validity of the proposed approach.

No abstract