Science – Page 78

Forecasting with Confidence: A Deep Dive into Prophet

13.01.2026 by qfx

Forecasting accuracy benefits from decomposing retail data into constituent parts - trend, yearly and weekly seasonality, and holiday effects - allowing for a nuanced understanding of predictive components.

This review examines the Prophet forecasting framework, highlighting its strengths in building reliable and understandable time series models.

The AI Plateau: Why Machines Can’t Teach Themselves to Think

13.01.2026 by qfx

A new analysis reveals fundamental limits to self-improvement in current AI systems, demonstrating why true artificial general intelligence remains a distant prospect.

Pricing Complexity: A New Approach to Derivative Valuation

13.01.2026 by qfx

$The optimal transport plan reveals a concentration of probability mass along a diagonal-consistent with the martingale constraint [latex]\mathbb{E}[X\_{1}|X\_{0}]=X\_{0}[/latex]-and highlights a high-probability transition path clustered near the point [latex](5500, 6500)[/latex].$

This review introduces a novel neural network framework for efficiently solving multi-period martingale optimal transport problems, accelerating the pricing and risk management of complex financial derivatives.

Decoding Software: From Bug Reports to Clear Explanations

12.01.2026 by qfx

Retrieval-Augmented Generation (RAG) establishes a synergistic framework where pre-trained language models leverage external knowledge sources to enhance response generation, effectively combining the strengths of parametric knowledge stored within the model with non-parametric knowledge retrieved from a database, thus improving both accuracy and adaptability.

A new approach leverages issue tracking data and advanced language models to automatically generate insightful explanations of software behavior, offering a powerful alternative to traditional documentation.

Mapping the Attack Path: How Graph Networks Are Reinventing Lateral Movement Detection

12.01.2026 by qfx

Large language model pretraining conventionally relies on predicting masked tokens within sampled sentences, but this work extends that paradigm by treating random walks through a graph as those sentences-where node IDs and edge features function as the tokens to be predicted-thereby framing the learning process as a navigation of relationships rather than a parsing of sequential text.

A new approach leverages the power of graph-based machine learning to identify malicious activity as it spreads within enterprise networks.

Mapping Cause and Effect with Enhanced Graph Learning

12.01.2026 by qfx

A new method improves the accuracy and efficiency of uncovering causal relationships by focusing on robustly learning the underlying structure of complex systems.

Beyond Tables: Semantic Data Boosts Transaction Analysis

12.01.2026 by qfx

The proposed method offers a comprehensive approach to the problem, laying the groundwork for future iterations despite the inevitable accumulation of technical debt inherent in any complex system.

New research demonstrates how incorporating the meaning of categorical data improves the performance of foundation models in understanding financial transactions.

Navigating Market Shifts: A Deep Learning Approach to Macro Portfolio Management

12.01.2026 by qfx

The DeePM pipeline processes asset histories through a hybrid backbone, then leverages causal directed delays to attend to global state, refining latent embeddings via a macro-graph GNN, all while minimizing a robust loss function combining pooled Net Sharpe ratio and a worst-window SoftMin penalty to navigate the chaos of financial modeling.

This research introduces a novel deep learning framework designed to enhance the resilience and performance of systematic macro trading strategies in dynamic economic environments.

Untangling Demand: From Raw Data to Reliable Insights

12.01.2026 by qfx

$LM1LM\_{1} provides a diagnostic capability when applied empirically.$

New research offers a robust toolkit for overcoming measurement challenges and building more accurate demand estimations using unstructured data.

Crisis Signals: How Investor Behavior Shifts When Markets Turn

12.01.2026 by qfx

The study demonstrates that while Kalman filtering offers a theoretically elegant approach to state estimation, practical implementation inevitably exposes limitations when confronted with real-world noise and system dynamics, ultimately necessitating careful consideration of its assumptions and potential for divergence.

New research reveals that foreign investor order flow becomes a much stronger predictor of market movements during times of crisis, demanding more dynamic modeling approaches.