Start with the subject "Strange" as your core object. Identify its "shallow" attributes—the immediate, literal qualities. Distorted, uncanny, or out-of-place. Emotional: Unsettling, curious, or eerie. 🧬 Step 2: Apply Feature Primitives
"Deep Feature Synthesis" suggests layering these primitives to find a "load-bearing fact" that changes perception. "It is a weird image."
If you were building this feature for an AI model or a narrative, you might define it as: Strange
Are you building a (technical) or a character/story (creative)?
A "deep" feature is built by applying operations to these base attributes across different dimensions, such as time, frequency, or relationship to "normalcy." Start with the subject "Strange" as your core object
For a visual example of how 'deep features' are identified in complex subjects like human facial structures, watch this explanation: Stop Drawing Noses Tony Swaby YouTube• Apr 23, 2026 If you want to refine this further, tell me:
This specific deep feature doesn't just say something is "strange"—it measures the between how real something looks and how wrong it feels over time. Emotional: Unsettling, curious, or eerie
What is the of "Strange"? (e.g., a physical object, a feeling, or a mathematical anomaly?)