Selected work

Representative Papers

Quant Convergence: Bridging Classical Value Investing and Modern Factor Models for Systematic Equity Selection

Jun 23, 2026

This study addresses the tendency of modern machine learning models to overfit short-term market noise in stock selection, thereby neglecting firms’ long-term intrinsic value. To bridge this gap, the authors formally operationalize Benjamin Graham’s “margin of safety” principle as a low-pass filter, integrating his classic value-investing rules with contemporary factor models to construct three distinct feature sets. Using two decades of S&P 500 data, they evaluate XGBoost, Random Forest, and AutoGluon under a four-year buy-and-hold strategy to assess out-of-sample robustness. Results demonstrate that a pure Graham-inspired Random Forest achieves a cumulative return of 232.13% and a Calmar ratio of 1.38, while the hybrid model attains a 202.91% return with the lowest drawdown (34.53%), significantly outperforming high-volatility purely AI-driven approaches.

0 citationsRead paper

Multimodal Image Colorization: Quantifying the Impact of Text-Conditioned Guidance on Grayscale-to-Color Translation

Jun 16, 2026

This work addresses the inherent ambiguity in grayscale image colorization, which often leads to semantically inaccurate color predictions. To mitigate this issue, the authors propose incorporating CLIP-based textual conditioning as a guidance signal and present the first systematic evaluation—under controlled conditions—of how text prompts influence colorization performance across two distinct architectures: a U-Net and Stable Diffusion 1.5. Experimental results demonstrate that text guidance substantially enhances colorization quality: for the U-Net, PSNR improves by 5.6%, SSIM by 1.2%, colorfulness by 36.6%, and LPIPS decreases by 7.6%. Consistent improvements are also observed with Stable Diffusion, confirming the effectiveness and generalizability of text-guided colorization across different model architectures.

0 citationsRead paper

Scribby: A Multi-Level LLM Framework for Semantic Video Analysis

Jun 08, 2026

This work addresses the limitations of existing video analysis methods, which rely on coarse-grained AI summaries and struggle to capture the structural evolution and semantic relationships in long-form videos. To overcome this, the authors propose a multi-level large language model (LLM) framework that first performs global semantic modeling over full video transcripts, then conducts context-aware sentence-by-sentence parsing. Innovatively, the framework employs an LLM as a discriminator to cluster sentences based on semantic similarity, enabling fine-grained semantic segmentation. By integrating both global and local contextual information, the approach supports the generation of interpretable visualizations—including semantic structure graphs and relevance heatmaps—thereby significantly enhancing the depth and explainability of video understanding. This method is particularly well-suited for applications such as educational content analysis and lecture replay, where structured semantic insight is essential.

0 citationsRead paper

AI Driven Soccer Analysis Using Computer Vision

Apr 09, 2026

This study addresses the challenge of accurately estimating player positions and deriving tactical metrics from broadcast soccer footage, which conventional video analysis methods struggle to achieve due to perspective distortions. To overcome this limitation, the authors propose an integrated framework combining object detection (YOLO/Faster R-CNN), instance segmentation (SAM2), and CNN-based keypoint localization. By leveraging homography transformation, the method maps players and field elements from arbitrary camera views into real-world coordinates, enabling high-fidelity reconstruction of player trajectories for the first time. The approach reliably computes tactical indicators—including player positions, velocities, distances covered, and heatmaps—providing coaches with actionable insights that are inaccessible through traditional analytical techniques.

0 citationsRead paper

SkyNet: Belief-Aware Planning for Partially-Observable Stochastic Games

Mar 29, 2026

This work addresses the challenge of modeling belief states in partially observable, stochastic multi-agent games, where existing approaches such as MuZero struggle to effectively represent hidden information. The authors propose a belief-aware mechanism that operates within the MuZero framework without requiring explicit belief tracking or modifications to the search algorithm. By introducing a self-conditioned auxiliary prediction head trained via supervisory signals from winner prediction and rank estimation, the latent state representation is encouraged to retain outcome-predictive information. Integrated with a Transformer encoder, Monte Carlo tree search, heuristic opponent curriculum learning, and self-play training, the method achieves a 75.3% win rate (equivalent to +194 Elo) against baseline agents in the card game Skyjo, and improves the win rate against heuristic opponents from 0.466 to 0.720, substantially outperforming prior approaches.

0 citationsRead paper

Recent publications

Latest Papers

Quant Convergence: Bridging Classical Value Investing and Modern Factor Models for Systematic Equity Selection

Jun 23, 2026

This study addresses the tendency of modern machine learning models to overfit short-term market noise in stock selection, thereby neglecting firms’ long-term intrinsic value. To bridge this gap, the authors formally operationalize Benjamin Graham’s “margin of safety” principle as a low-pass filter, integrating his classic value-investing rules with contemporary factor models to construct three distinct feature sets. Using two decades of S&P 500 data, they evaluate XGBoost, Random Forest, and AutoGluon under a four-year buy-and-hold strategy to assess out-of-sample robustness. Results demonstrate that a pure Graham-inspired Random Forest achieves a cumulative return of 232.13% and a Calmar ratio of 1.38, while the hybrid model attains a 202.91% return with the lowest drawdown (34.53%), significantly outperforming high-volatility purely AI-driven approaches.

0 citationsRead paper

Multimodal Image Colorization: Quantifying the Impact of Text-Conditioned Guidance on Grayscale-to-Color Translation

Jun 16, 2026

This work addresses the inherent ambiguity in grayscale image colorization, which often leads to semantically inaccurate color predictions. To mitigate this issue, the authors propose incorporating CLIP-based textual conditioning as a guidance signal and present the first systematic evaluation—under controlled conditions—of how text prompts influence colorization performance across two distinct architectures: a U-Net and Stable Diffusion 1.5. Experimental results demonstrate that text guidance substantially enhances colorization quality: for the U-Net, PSNR improves by 5.6%, SSIM by 1.2%, colorfulness by 36.6%, and LPIPS decreases by 7.6%. Consistent improvements are also observed with Stable Diffusion, confirming the effectiveness and generalizability of text-guided colorization across different model architectures.

0 citationsRead paper

Scribby: A Multi-Level LLM Framework for Semantic Video Analysis

Jun 08, 2026

This work addresses the limitations of existing video analysis methods, which rely on coarse-grained AI summaries and struggle to capture the structural evolution and semantic relationships in long-form videos. To overcome this, the authors propose a multi-level large language model (LLM) framework that first performs global semantic modeling over full video transcripts, then conducts context-aware sentence-by-sentence parsing. Innovatively, the framework employs an LLM as a discriminator to cluster sentences based on semantic similarity, enabling fine-grained semantic segmentation. By integrating both global and local contextual information, the approach supports the generation of interpretable visualizations—including semantic structure graphs and relevance heatmaps—thereby significantly enhancing the depth and explainability of video understanding. This method is particularly well-suited for applications such as educational content analysis and lecture replay, where structured semantic insight is essential.

0 citationsRead paper

AI Driven Soccer Analysis Using Computer Vision

Apr 09, 2026

This study addresses the challenge of accurately estimating player positions and deriving tactical metrics from broadcast soccer footage, which conventional video analysis methods struggle to achieve due to perspective distortions. To overcome this limitation, the authors propose an integrated framework combining object detection (YOLO/Faster R-CNN), instance segmentation (SAM2), and CNN-based keypoint localization. By leveraging homography transformation, the method maps players and field elements from arbitrary camera views into real-world coordinates, enabling high-fidelity reconstruction of player trajectories for the first time. The approach reliably computes tactical indicators—including player positions, velocities, distances covered, and heatmaps—providing coaches with actionable insights that are inaccessible through traditional analytical techniques.

0 citationsRead paper

SkyNet: Belief-Aware Planning for Partially-Observable Stochastic Games

Mar 29, 2026

This work addresses the challenge of modeling belief states in partially observable, stochastic multi-agent games, where existing approaches such as MuZero struggle to effectively represent hidden information. The authors propose a belief-aware mechanism that operates within the MuZero framework without requiring explicit belief tracking or modifications to the search algorithm. By introducing a self-conditioned auxiliary prediction head trained via supervisory signals from winner prediction and rank estimation, the latent state representation is encouraged to retain outcome-predictive information. Integrated with a Transformer encoder, Monte Carlo tree search, heuristic opponent curriculum learning, and self-play training, the method achieves a 75.3% win rate (equivalent to +194 Elo) against baseline agents in the card game Skyjo, and improves the win rate against heuristic opponents from 0.466 to 0.720, substantially outperforming prior approaches.

0 citationsRead paper