I am unable to create images. However, I can provide a detailed description to guide the creation of an "IEEE TIFS paper quality" system framework diagram (Framework Overview Figure) on the theme of: DENIA: Budget-Adaptive Black-Box Node Injection against GNN-Based Malware Detection Overall Requirements: * Purpose: IEEE TIFS paper Fig. 1 (ready for submission), requiring clarity, professionalism, vector style, and printability. * Canvas: Landscape orientation, pure white background, aspect ratio approximately 3.2:1 (close to 16:5), with sufficient margins. * Style: Flat vector illustration, no gradients, no shadows, no noise textures, clean lines. * Font: Sans-serif (similar to Helvetica/Arial), primarily English, mathematical symbols in LaTeX style appearance (Δ, λ, α, ℝ, Σ, etc.). * Lines: Uniform line width; node circles with black outlines; edges as thin black lines; process arrows as thicker black arrows. * Structure: Three-stage pipeline from left to right, with three main modules in parallel, connected by thick black arrows (→). * Each large module uses a rounded rectangle outline with a light background and dark outline, with the module title centered at the top. * A large title is written centered at the bottom of the entire diagram (must be clearly readable): DENIA: Budget-Adaptive Black-Box Node Injection against GNN-Based Malware Detection Global Title Description (placed as small text at the top of the entire image, in light gray): * Add a line of small, light gray explanatory text at the very top of the image (without obscuring the main body): "Strict black-box setting: only predicted class probabilities are accessible (no gradients, no parameters)." Module ①: Dynamic Budget Allocation (left side, light blue background box) 1. 1 Module Outline and Title * The first large rounded box on the left, with a light blue background and blue outline. * Top title (bold): Dynamic Budget Allocation 1. 2 Sub-diagram: Original Malware CFG (top left area) * Label: Original Malware CFG * Draw an undirected CFG graph: 10~14 light gray nodes (circular, black outline), black edges connecting to form a reasonably structured graph. * Select 2~3 "target nodes" filled with light blue and highlighted with dashed circles. * Next to it
Figure 1: Risk Types in Deep Application of Artificial Intel...