Update README.md

README.md

@@ -96,6 +96,27 @@ configs:
       path: visual_prompting_pairs/visual_prompting_val.parquet
 ---
 
+Motivation: A key question for understanding multimodal performance is analyzing the ability for a model to have basic vs. detailed understanding of images. These capabilities are needed for models to be used in
+real-world tasks, such as an assistant in the physical world. While there are many dataset for object detection
+and recognition, there are few that test spatial reasoning and other more targeted task such as visual prompting.
+The datasets that do exist are static and publicly available, thus there is concern that current AI models could
+be trained on these datasets, which makes evaluation with them unreliable. Thus we created a dataset that is
+procedurally generated and synthetic, and tests spatial reasoning, visual prompting, as well as object recognition
+and detection [91] . The datasets are challenging for most AI models and by being procedurally generated the
+16
+benchmark can be regenerated ad infinitum to create new test sets to combat the effects of models being trained
+on this data and the results being due to memorization.
+Benchmark Description: This dataset has 4 sub-tasks: Object Recognition, Visual Prompting. Spatial Rea-
+soning, and Object Detection. For each sub-task, the images consist of images of pasted objects on random
+images. The objects are from the COCO [62] object list and are gathered from internet data. Each object is
+masked using the DeepLabV3 object detection model [22] and then pasted on a random background from the
+Places365 dataset [132]. The objects are pasted in one of four locations, top, left, bottom, and right, with small
+amounts of random rotation, positional jitter, and scale.
+There are 2 conditions “ single” and “ pairs”, for images with one and two objects. Each test set uses 20
+sets of object classes (either 20 single objects or 20 pairs of objects), with four potential locations and four
+backgrounds classes, and we sample 4 instances of object and background. This results in 1280 images per
+condition and sub-task.
+
 What are the experimental design setup dimensions 
 (e.g. settings, prompt templates, dataset subsets) for this benchmark?