Conference Logo ProgramMaster Logo
Conference Tools for TMS Specialty Congress 2026
Login
Register as a New User
Help
Submit An Abstract
Propose A Symposium
Presenter/Author Tools
Organizer/Editor Tools

About this Abstract
Meeting TMS Specialty Congress 2026
Symposium 4th World Congress on Artificial Intelligence in Materials & Manufacturing (AIM 2026)
Presentation Title A Generalized Machine Learning Framework for Data-Driven Prediction of Relative Density in Laser Powder Bed Fusion Parts
Author(s) Abdul Khalad, Kondababu Kadali, Gururaj Telasang, Peng Zhang, Wei Xu, Viswanath Chinthapenta
On-Site Speaker (Planned) Abdul Khalad
Abstract Scope Achieving high relative density (RD) is critical for new alloy systems fabricated using laser powder bed fusion (L-PBF). However, the conventional design of experiments (DOE) becomes inefficient due to the large number of interacting process parameters. This study presents a data-driven machine learning (ML) framework to confine the optimization search space to the most influential parameters. A dataset compiled from literature over the past decade, covering 11 alloy systems, was divided into 80:20 training and testing sets. Multiple ML models were evaluated to capture nonlinear process–property relationships. The hybrid gradient boosting–particle swarm optimization (GB-PSO) model demonstrated superior performance, achieving MAE/R² values of 0.20/0.99 (training) and 0.73/0.95 (testing). SHAP analysis provided insight into global and local parameter significance. The reduced experimental design derived from the ML framework was validated through in-house fabrication of Inconel 718, confirming its effectiveness in achieving high RD.
Proceedings Inclusion? Undecided

OTHER PAPERS PLANNED FOR THIS SYMPOSIUM

3D PRIMME for Learning Grain Growth Behaviors from Simulated and Experimental Datasets
A Generalized Machine Learning Framework for Data-Driven Prediction of Relative Density in Laser Powder Bed Fusion Parts
A Reinforcement Learning Approach to Alloy Design for Energy Materials
Accelerated Prediction of EAF Melting Processes Through an Xgboost-Based Reduced-Order Model
Accelerating Design-Plan-Print-Qualify Workflows for Additive Manufacturing with Agentic AI
Adaption to Uncontrolled Variables in Additive Manufacturing
AFM-net: From Scarce Data to Fast Scans. Machine Learning Acceleration of AFM Nanometrology
AI-Driven Alloy Design Framework for Cost-Effective Ultrahigh Strength Martensitic Steels
AI-Driven Raman–Terahertz System for Real-Time Measurement of Polymer Crystallinity
AI-Enabled Microstructure Characterization of an Al–Si–Mg–Cu Die-Casting Alloy Under 490 °C Solution Treatment
ANN Based Modeling of Thermodynamic, Transport, and Transformation Related Properties
Automated Analysis Using Deep Learning for Complex Low-Temperature Transformation Microstructures in Advanced High-Strength Steel
Beyond Trial and Error: Bayesian Paradigms for Intelligent Alloy Design
Chemical-Disorder-Informed Machine Learning Framework for Multi-Objective Design of Refractory Multi-Principal Element Alloys
Data-Driven Discovery of Creep Equations via Symbolic Regression
Diagnosing Mechanical Assembly Discontinuities with LLMs
Empower Finite Element Software to Perform Machine Learning for Modelling Material Behaviours
European Infrastructure Activities on Management of Materials Data for Research, Development and Application of Advanced Materials
Evaluating and Optimizing Large Language Models in Manufacturing Contexts
Extracting Symbolic Relationships in Materials Science and Engineering
From ICME to Industry: A Platform Approach to the Discovery and Commercialization of Advanced Alloys
From Quantum-Informed Descriptors to Additive Manufacturing Process: A Scale-Bridging Workflow for TCP Phase Prediction
Generalist to Specialist Sequential Segmentation of Label-Free Low-Contrast Al-Si Solidification Video
High-Fidelity Phase-Field lattice Boltzmann Simulation for Dataset Generation Enabling Machine Learning of Melt Pool Dynamics in PBF
ICME-Driven, Uncertainty-Aware Surrogate Modeling of Composite EV Battery Enclosures Using High-Throughput Simulations and Machine Learning
In-Situ Prediction of 3D-Printed Outcomes
Integrated Generative and Predictive Machine Learning Models for Estimation of Microstructure and Cold Dwell Fatigue Life in Ti-6Al-4V Forgings
Integrated Machine Learning Framework for Quality Prediction of Hot Forging
Leveraging Large Language Models for Alloy Property Prediction: A Review of Regression Approaches
Machine Learning-Assisted Design of Resorbable Magnesium Alloys
Machine Learning-Based Prediction of Mechanical Property Heterogeneity in Inconel 718 Superalloy Manufactured by Directed Energy Deposition
Microstructure-Sensitive Segmentation of γ′ Phase in Superalloys Using Detectron2 for Data-Driven Alloy Design and Process Optimisation
Morphology-Based Estimation of Material Parameters via Physics-Informed Neural Networks
Neural Modeling of Anisotropic Yield Behavior Under Plasticity Constraints
Optimal Resource Utilization for Autonomous Lab Orchestrators
Phsyics-Guided Machine Learning for Predicting Multiscale Failure in Composites
Physics-Guided 1D Convolutional Neural Networks for Scalable Prediction of Thermo-Mechanical Response in Laser Powder Bed Fusion
Real-Time Digital Twin for EAF Process Optimization
Semantic Knowledge Graphs for AI-Ready Data
Simulation-Driven Machine Learning for Predicting Material Properties from Cross-Sectional Dendritic Microstructures
Superintelligence for Scientific Discovery: Multi-Agent Swarms and Large Reasoning Models
Towards a Modular Autonomous Research Ecosystem for Materials Science
Unlocking the Black Box: A Neuro-Symbolic Approach to Modeling TCP Phase Precipitation in Sustainable High-Performance Alloys
Unsupervised Microstructure Segmentation of Forged TiAl Alloys Using Machine Learning
Update-Free Reinforcement Learning for Scalable Resilience in Expeditionary Conditions
μSAM-Based Inference of Nuclear Materials Processing History from SEM Imagery
Questions about ProgramMaster? Contact programming@programmaster.org | TMS Privacy Policy | Accessibility Statement