Real-world decision-making often requires integrating and reasoning over information from multiple modalities. While recent multimodal large language models (MLLMs) have shown promise in such tasks, their ability to perform multi-hop reasoning across diverse sources remains insufficiently evaluated. Existing benchmarks, such as MMQA, face challenges due to (1) data contamination and (2) a lack of complex queries that necessitate operations across more than two modalities, hindering accurate performance assessment. To address this, we present Financial Cross-Modal Multi-Hop Reasoning (FCMR), a benchmark created to analyze the reasoning capabilities of MLLMs by urging them to combine information from textual reports, tables, and charts within the financial domain. FCMR is categorized into three difficulty levels—Easy, Medium, and Hard—facilitating a step-by-step evaluation. In particular, problems at the Hard level require precise cross-modal three-hop reasoning and are designed to prevent the disregard of any modality. Experiments on this new benchmark reveal that even state-of-the-art MLLMs struggle, with the best-performing model (Claude 3.5 Sonnet) achieving only 30.4% accuracy on the most challenging tier. We also conduct analysis to provide insights into the inner workings of the models, including the discovery of a critical bottleneck in the information retrieval phase.

This paper considers a dynamic spectrum allocation (DSA) model that a spectrum broker (SB) coordinates the allocation of the spectrum with the regional license inside the region of responsibility. It has a potential to leverage the spectrum utilization but requires a sophisticated approach to manage the allocation-dependent interference effect. Thus, the SB is responsible to manage the wireless interference between base stations (BSs) within the region for quality of service (QoS) provisioning over the allocated channels. In this paper, we address the interference constrained DSA problem and propose an interference management scheme that collaboratively works with the spectrum allocation algorithm to implement the DSA. By the allocation-aware interference management based on the received signal-to-interference-plus-noise-ratio (SINR), the proposed scheme can reflect the context of the SB's decisions at the allocation process. Simulation results demonstrate that the proposed scheme efficiently distributes the spectrum resource while guaranteeing the QoS requirements of all BSs with allocated channels.