arXiv subjects

Parth Arora

Publications and source records attributed to Parth Arora.

3 recordsLinked to original sources

arXiv2026

Evaluating Financial Intelligence in Large Language Models: Benchmarking SuperInvesting AI with LLM Engines

Large language models are increasingly used for financial analysis and investment research, yet systematic evaluation of their financial reasoning capabilities remains limited. In this work, we introduce the AI Financial Intelligence Benchmark (AFIB), a multi-dimensional evaluation framework designed to assess financial analysis capabilities across five dimensions: factual accuracy, analytical completeness, data recency, model consistency, and failure patterns. We evaluate five AI systems: GPT, Gemini, Perplexity, Claude, and SuperInvesting, using a dataset of 95+ structured financial analysis questions derived from real-world equity research tasks. The results reveal substantial differences in performance across models. Within this benchmark setting, SuperInvesting achieves the highest aggregate performance, with an average factual accuracy score of 8.96/10 and the highest completeness score of 56.65/70, while also demonstrating the lowest hallucination rate among evaluated systems. Retrieval-oriented systems such as Perplexity perform strongly on data recency tasks due to live information access but exhibit weaker analytical synthesis and consistency. Overall, the results highlight that financial intelligence in large language models is inherently multi-dimensional, and systems that combine structured financial data access with analytical reasoning capabilities provide the most reliable performance for complex investment research workflows.

cs.AI↗

arXiv2025

Positioning Monocular Optical See Through Head Worn Displays in Glasses for Everyday Wear

Head-worn displays for everyday wear in the form of regular eyeglasses are technically feasible with recent advances in waveguide technology. One major design decision is determining where in the user's visual field to position the display. Centering the display in the principal point of gaze (PPOG) allows the user to switch attentional focus between the virtual and real images quickly, and best performance often occurs when the display is centered in PPOG or is centered vertically below PPOG. However, these positions are often undesirable in that they are considered interruptive or are associated with negative social perceptions by users. Offsetting the virtual image may be preferred when tasks involve driving, walking, or social interaction. This paper consolidates findings from recent studies on monocular optical see-through HWDs (OST-HWDs), focusing on potential for interruption, comfort, performance, and social perception. For text-based tasks, which serve as a proxy for many monocular OST-HWD tasks, we recommend a 15{\deg} horizontal field of view (FOV) with the virtual image in the right lens vertically centered but offset to +8.7{\deg} to +23.7{\deg} toward the ear. Glanceable content can be offset up to +30{\deg} for short interactions.

cs.HC↗

arXiv2025

A Novel SIMD-Optimized Implementation for Fast and Memory-Efficient Trigonometric Computation

This paper proposes a novel set of trigonometric implementations which are 5x faster than the inbuilt C++ functions. The proposed implementation is also highly memory efficient requiring no precomputations of any kind. Benchmark comparisons are done versus inbuilt functions and an optimized taylor implementation. Further, device usage estimates are also obtained, showing significant hardware usage reduction compared to inbuilt functions. This improvement could be particularly useful for low-end FPGAs or other resource-constrained devices.

cs.MS↗