Research & Analysis

In-depth research papers, case studies, surveys, and academic insights on finance and technology

Articles

356

Updated

April 23, 2026

When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs

Section 1 – What happened? Swiss fintech companies are grappling with a pressing issue in the development of large vision-language models (LVLMs): the…

Research & Analysis

When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs

Top Stories

The Sample Complexity of Multicalibration

Temporal Taskification in Streaming Continual Learning: A Source of Evaluation Instability

Seeing Fast and Slow: Learning the Flow of Time in Videos

Fine-Tuning Regimes Define Distinct Continual Learning Problems

More Stories

When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs

The Sample Complexity of Multicalibration

Temporal Taskification in Streaming Continual Learning: A Source of Evaluation Instability

Seeing Fast and Slow: Learning the Flow of Time in Videos

Fine-Tuning Regimes Define Distinct Continual Learning Problems

SpeechParaling-Bench: A Comprehensive Benchmark for Paralinguistic-Aware Speech Generation

Global Offshore Wind Infrastructure: Deployment and Operational Dynamics from Dense Sentinel-1 Time Series

FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels

Closing the Domain Gap in Biomedical Imaging by In-Context Control Samples

AVISE: Framework for Evaluating the Security of AI Systems

SpeechParaling-Bench: A Comprehensive Benchmark for Paralinguistic-Aware Speech Generation

Global Offshore Wind Infrastructure: Deployment and Operational Dynamics from Dense Sentinel-1 Time Series

FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels

Closing the Domain Gap in Biomedical Imaging by In-Context Control Samples

AVISE: Framework for Evaluating the Security of AI Systems

UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling

Safe Continual Reinforcement Learning in Non-stationary Environments

Phase Transitions in the Fluctuations of Functionals of Random Neural Networks

Generalization at the Edge of Stability

FASTER: Value-Guided Sampling for Fast RL

UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling

Safe Continual Reinforcement Learning in Non-stationary Environments

Phase Transitions in the Fluctuations of Functionals of Random Neural Networks

Generalization at the Edge of Stability

FASTER: Value-Guided Sampling for Fast RL

When Can LLMs Learn to Reason with Weak Supervision?

Sessa: Selective State Space Attention

MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval

Bounded Ratio Reinforcement Learning

Agentic Forecasting using Sequential Bayesian Updating of Linguistic Beliefs

When Can LLMs Learn to Reason with Weak Supervision?

Sessa: Selective State Space Attention

MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval

Bounded Ratio Reinforcement Learning

Agentic Forecasting using Sequential Bayesian Updating of Linguistic Beliefs

MM-WebAgent: A Hierarchical Multimodal Web Agent for Webpage Generation

How Do LLMs and VLMs Understand Viewpoint Rotation Without Vision? An Interpretability Study

Generalization in LLM Problem Solving: The Case of the Shortest Path

Diagnosing LLM Judge Reliability: Conformal Prediction Sets and Transitivity Violations

Benchmarking Optimizers for MLPs in Tabular Deep Learning

MM-WebAgent: A Hierarchical Multimodal Web Agent for Webpage Generation

How Do LLMs and VLMs Understand Viewpoint Rotation Without Vision? An Interpretability Study

Generalization in LLM Problem Solving: The Case of the Shortest Path

Diagnosing LLM Judge Reliability: Conformal Prediction Sets and Transitivity Violations

Benchmarking Optimizers for MLPs in Tabular Deep Learning

Rhetorical Questions in LLM Representations: A Linear Probing Study

LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoning

HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation System

From $P(y|x)$ to $P(y)$: Investigating Reinforcement Learning in Pre-train Space

From Feelings to Metrics: Understanding and Formalizing How Users Vibe-Test LLMs

Rhetorical Questions in LLM Representations: A Linear Probing Study

LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoning

HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation System

From $P(y|x)$ to $P(y)$: Investigating Reinforcement Learning in Pre-train Space

From Feelings to Metrics: Understanding and Formalizing How Users Vibe-Test LLMs

Visual Preference Optimization with Rubric Rewards

Representation geometry shapes task performance in vision-language modeling for CT enterography

PAL: Personal Adaptive Learner

Classical and Quantum Speedups for Non-Convex Optimization via Energy Conserving Descent

CLAD: Efficient Log Anomaly Detection Directly on Compressed Representations

Visual Preference Optimization with Rubric Rewards

Representation geometry shapes task performance in vision-language modeling for CT enterography

PAL: Personal Adaptive Learner

Classical and Quantum Speedups for Non-Convex Optimization via Energy Conserving Descent

CLAD: Efficient Log Anomaly Detection Directly on Compressed Representations

Solving Physics Olympiad via Reinforcement Learning on Physics Simulators

Physics-Informed State Space Models for Reliable Solar Irradiance Forecasting in Off-Grid Systems

Detecting Safety Violations Across Many Agent Traces

C-ReD: A Comprehensive Chinese Benchmark for AI-Generated Text Detection Derived from Real-World Prompts

Budget-Aware Uncertainty for Radiotherapy Segmentation QA Using nnU-Net

Solving Physics Olympiad via Reinforcement Learning on Physics Simulators

Physics-Informed State Space Models for Reliable Solar Irradiance Forecasting in Off-Grid Systems