Given the vast amount of unstructured financial text data available today, there is a high demand for reliable, quality annotations to facilitate robust model development. However, traditional methods can often be expensive and time-inefficient. In this study, we investigate annotations for emotion, sentiment, and cognitive dissonance generated by the large language models (LLMs), GPT-3.5 and GPT-4, for quarterly earnings conference calls and compare them against human annotations obtained via traditional methods. We also investigate different prompt engineering choices on LLM annotation quality, experimenting with 4 styles of prompts centered around varying the amount of contextual information given and how it is presented to the models. Our results show the GPT models are not only more consistent and reliable than human annotators, but also provide annotations in a more cost- and time-efficient manner.

We introduce sSTNN, a sparse and parallelized version of a spatio-temporal neural network (STNN) that enables training on much larger datasets. First, we introduce the model architecture and discuss the modifications we made to enable the use of a sparse data structure and multi-GPU parallelization. Then we present empirical results that demonstrate sSTNNs ability to train and inference on a dataset 17 times larger than STNN is capable of. Finally, we discuss the effect of sparsification on runtime and present evidence that sSTNN can achieve upwards of 117× reduction in memory usage compared to STNN.

A significant amount of resources have been used in both academia and industry to study the impact of financial text on company perception and performance. In order to mitigate potential adverse outcomes, companies have begun to regulate word usage based on perceived sentiment, making conventional text-based analysis less reliable. To address this, we present a multimodal bidirectional Long Short-Term Memory (LSTM) framework augmented with a cross-attention fusion mechanism trained on audio and text data obtained from quarterly earnings conferences calls. The framework is applied to two tasks: financial restatement prediction and market movement prediction. We compare the proposed model against several baseline methods and find that while it does not achieve superior performance, our results show that utilizing multimodal data leads to a substantial increase in model accuracy for restatement prediction. Furthermore, we gain insight on the effectiveness of semantic-and emotion-related features towards these tasks.

We present an extension to the model-free anomaly detection algorithm, Isolation Forest. This extension, named Extended Isolation Forest (EIF), resolves issues with assignment of anomaly score to given data points. We motivate the problem using heat maps for anomaly scores. These maps suffer from artifacts generated by the criteria for branching operation of the binary tree. We explain this problem in detail and demonstrate the mechanism by which it occurs visually. We then propose two different approaches for improving the situation. First we propose transforming the data randomly before creation of each tree, which results in averaging out the bias. Second, which is the preferred way, is to allow the slicing of the data to use hyperplanes with random slopes. This approach results in remedying the artifact seen in the anomaly score heat maps. We show that the robustness of the algorithm is much improved using this method by looking at the variance of scores of data points distributed along constant level sets. We report AUROC and AUPRC for our synthetic datasets, along with real-world benchmark datasets. We find no appreciable difference in the rate of convergence nor in computation time between the standard Isolation Forest and EIF.

The next-generation astronomy digital archives will cover most of the sky at fine resolution in many wavelengths, from X-rays, through ultraviolet, optical, and infrared. The archives will be stored at diverse geographical locations. One of the first of these projects, the Sloan Digital Sky Survey (SDSS) is creating a 5-wavelength catalog over 10,000 square degrees of the sky (see http://www.sdss.org/). The 200 million objects in the multi-terabyte database will have mostly numerical attributes in a 100+ dimensional space. Points in this space have highly correlated distributions.