**Solving a Temporal Graph Neural Network (TGNN) Challenge f

#ai #compliance #pld

Solving a Temporal Graph Neural Network (TGNN) Challenge for Real-Time Traffic Forecasting

Imagine you are tasked with developing a temporal graph neural network to predict the traffic congestion level in a city's road network over the next hour, given the current real-time traffic data, weather conditions, and time of day. The twist: you must incorporate both spatial and temporal graph structures into your model, as well as account for the presence of periodic events like rush hour, festivals, and construction roadblocks.

Constraints:

Graph Size: The road network is composed of 1000 nodes (intersections) with an average of 200 edges (roads connecting intersections), resulting in a dense graph with 200,000 edges.
Temporal Resolution: You have access to 1-minute resolution traffic data for the past 24 hours, which you will use to train your model.
Weather Data: You have access to real-time weather data including temperature, humidity, wind speed, and precipitation, which you must incorporate into your model.
Model Evaluation: You will evaluate your model using a combination of mean absolute error (MAE), mean squared error (MSE), and area under the receiver operating characteristic curve (AUROC) metrics.
Computational Limitations: You are restricted to training your model on a single NVIDIA Tesla V100 GPU with 16 GB of memory, and a maximum of 4 hours of training time.

Objective:

Develop a temporal graph neural network that can accurately predict the traffic congestion level at each intersection in the road network over the next hour, given the real-time data and weather conditions.

Submission Requirements:

A well-documented Python codebase using TensorFlow or PyTorch as the deep learning framework.
A detailed description of your model architecture, including any novel graph neural network operations or techniques employed.
A plot of your model's performance on the evaluation metrics (MAE, MSE, AUROC).

Evaluation Criteria: