docs.monadam.dev

Bongard Problems

Bongard Problems require a combination of visual processing, abstraction, and reasoning, traits closely associated with human intelligence. As a measure of artificial general intelligence (AGI), Bongard Problems assess the ability of AI systems to generalize knowledge across different domains and scenarios. Python import cv2 import numpy as np from sklearn.ensemble import RandomForestClassifier # and preprocess images def preprocess_image(image_path): image = cv2.imread(image_path) image_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) return cv2.resize(image_gray, (64, 64)).flatten() # of loading images and extracting features images = ['image1.

Fargtown

Fargtown is a project for consolidating various implementations of ideas formulated by the Fluid Analogies Research Group (FARG). Copycat Melanie Mitchell’s Copycat is artificial intelligence software that focuses on analogical reasoning. Copycat can function with incomplete information, reflecting how humans can still find analogies even with partial data. It processes multiple hypotheses at once rather than sequentially. Copycat employs a fluid representation of concepts, adjusting its understanding based on the context and interactions.

Test Time Training

Cursory analysis of a test time training paper and the corresponding python code. Introduction The core idea of test time training (TTT) is finetuning an LM upon encountering previously-unseen input data. Definition Test time training updating model parameters temporarily during inference using a loss derived from input data. Results Over double the accuracy of fully-neural approaches (from 25% to 53%). When combined with program synthesis, this increases to 62%. For context, 62% accuracy is roughly average human performance; expert human performance is about 93%.