LLM-SR: Scientific Equation Discovery via Programming with Large Language Models

1 code implementation • 29 Apr 2024 • Parshin Shojaee, Kazem Meidani, Shashank Gupta, Amir Barati Farimani, Chandan K Reddy

Mathematical equations have been unreasonably effective in describing complex natural phenomena across various scientific disciplines.

SNIP: Bridging Mathematical Symbolic and Numeric Realms with Unified Pre-training

2 code implementations • 3 Oct 2023 • Kazem Meidani, Parshin Shojaee, Chandan K. Reddy, Amir Barati Farimani

To bridge the gap, we introduce SNIP, a Symbolic-Numeric Integrated Pre-training model, which employs contrastive learning between symbolic and numeric domains, enhancing their mutual similarities in the embeddings.

Contrastive Learning Few-Shot Learning +4

Graph-based Multi-ODE Neural Networks for Spatio-Temporal Traffic Forecasting

1 code implementation • 30 May 2023 • Zibo Liu, Parshin Shojaee, Chandan K Reddy

Current works primarily rely on road networks with graph structures and learn representations using graph neural networks (GNNs), but this approach suffers from over-smoothing problem in deep architectures.

Spatio-Temporal Forecasting Time Series Forecasting +1

Transformer-based Planning for Symbolic Regression

1 code implementation • NeurIPS 2023 • Parshin Shojaee, Kazem Meidani, Amir Barati Farimani, Chandan K. Reddy

Unlike conventional decoding strategies, TPSR enables the integration of non-differentiable feedback, such as fitting accuracy and complexity, as external sources of knowledge into the transformer-based equation generation process.

regression Symbolic Regression +1

Execution-based Code Generation using Deep Reinforcement Learning

1 code implementation • 31 Jan 2023 • Parshin Shojaee, Aneesh Jain, Sindhu Tipirneni, Chandan K. Reddy

It's important to note that PPOCoder is a task-agnostic and model-agnostic framework that can be used across different code generation tasks and PLs.

Code Completion Code Translation +5

Adaptively Weighted Top-N Recommendation for Organ Matching

no code implementations • 23 Jul 2021 • Parshin Shojaee, Xiaoyu Chen, Ran Jin

Because of the shortage, organ matching decision is the most critical decision to assign the limited viable organs to the most suitable patients.

Decision Making

Planar Maximum Coverage Location Problem with Partial Coverage, Continuous Spatial Demand, and Adjustable Quality of Service

no code implementations • 16 Dec 2020 • Manish Bansal, Parshin Shojaee

We consider a generalization of the classical planar maximum coverage location problem (PMCLP) in which partial coverage is allowed, facilities have adjustable quality of service (QoS) or service range, and demand zones and service zone of each facility are represented by two-dimensional spatial objects such as rectangles, circles, polygons, etc.

Optimization and Control Data Structures and Algorithms