Measuring dietary exposure is the key aspect of nutritional epidemiology in order to find cause and effect relationships between nutrition and long-term illnesses. Nevertheless, self-reported nutrition assessment tools e.g. food frequency questionnaires and dietary recalls provide systematic underreporting as well as random error in nutrition assessment that considerably reduce the regression coefficients of exposure-outcome relationships and even obscure true diet-health effects under measurement errors. The current corrections methods conducted on small reference samples and depending on the assumptions of linearity are capable of treating variations of errors in multimodal data of great dimensions. We are going to present an idea of self-supervised multimodal representation learning, that is, an error-reducting dietary exposure measure, where dietary text logs and wearable sensor data are modeled jointly and learns discriminative features highly correlated with true intake through cross-modal contrastive learning and masked reconstruction, trained over multi-view representations to produce an exposure-corrected dietary energy and nutrient consumption estimate using a unified latent space, and produce an exposure-corrected dietary text log estimate using a unified latent space.

Show more

With the widespread deployment of machine learning models in high-stakes decision-making contexts, their inherent opacity—often termed the "black-box" problem—has raised significant concerns regarding interpretability and reliability. This paper presents a systematic and comprehensive literature review examining the convergence of interpretable machine learning and statistical inference. This paper synthesizes foundational concepts, methodological frameworks, theoretical advancements, and practical applications to elucidate how statistical tools can validate, enhance, and formalize machine learning explanations. This review critically analyzes widely adopted techniques such as SHAP and LIME, and explores their integration with statistical inference tools, including hypothesis testing, confidence intervals, Bayesian methods, and causal inference frameworks. The analysis reveals that integrated approaches significantly improve explanation credibility, regulatory compliance, and decision transparency in critical domains, including healthcare diagnostics, financial risk management, and algorithmic governance. However, persistent challenges remain in theoretical consistency, computational efficiency, evaluation standardization, and human-centered design. This paper concludes by proposing a structured research agenda focusing on unified theoretical frameworks, efficient algorithmic implementations, domain-specific evaluation standards, and interdisciplinary collaboration strategies to advance the responsible development and deployment of explainable AI systems.

Show more

Field of view (FOV) algorithms are essential in determining the visible area of a player in 2D games. These algorithms dynamically calculate the visible areas while occluding these hidden areas, and play an important role in games such as roguelikes and stealth games. This survey summarizes three 2D FOV algorithms: ray casting, rectangle-based FOV, and recursive shadowcasting. The ray casting algorithm casts rays to determine which area was hidden from the player, which is a basic FOV algorithm. Rectangle-based FOV optimizes computation for large 2D grids by representing obstacles as rectangles, also using a quadtree to improve the access speed. Recursive shadowcasting efficiently computes the visible area by dividing the grid into 8 octants and recursively splitting the view when obstacles are encountered. This survey also mentioned how to adapt the recursive shadowcasting algorithm to 2.5D and 3D environments.

Show more

The well-known Large Language Model (LLM) compression is essential for enhancing computational efficiency, yet a systematic summary of investigation into structured pruning and low-rank decomposition remains absent in current literature. This work addresses the gap by providing a comprehensive review specifically focused on these two methodologies. Representative approaches are categorized and evaluated, including LLM-Pruner and SlimGPT for structured pruning, and ASVD and SVD-LLM for decomposition. These methods are rigorously analyzed in terms of algorithmic design, accuracy retention, and hardware adaptability. Through unified evaluation and comparative analysis, DISP-LLM and MoDeGPT are identified as the current state-of-the-art within their respective fields. Consequently, a conceptual framework is established to provide practical guidance for future research into efficient, training-free, and scalable LLM compression.

Show more