• Apache Commons, is an open-source for creating reusable Java components. It has numerical packages for linear algebra and non-linear optimization.
• ND4J is an open-source library that supports n-dimensional (ND) arrays, similar to NumPy. It runs on distributed GPUs or CPUs cross-platform. It supports the deep learning library Deeplearning4j.
• Colt provides a set of Open Source Libraries for High Performance Scientific and Technical Computing.
• Efficient Java Matrix Library (EJML) is an open-source linear algebra library for manipulating dense matrices.
• JAMA, a numerical linear algebra toolkit for the Java programming language. No active development has taken place since 2005, but it still one of the more popular linear algebra packages in Java.
• Jblas: Linear Algebra for Java, a linear algebra library which is an easy to use wrapper around BLAS and LAPACK.
• Parallel Colt is an open source library for scientific computing. A parallel extension of Colt.
• DataMelt, an open-source Java libraries for numerical calculations, data I/O and visualization of scientific results. Used together with Jython.
• Matrix Toolkit Java is a linear algebra library based on BLAS and LAPACK.
• OjAlgo is an open source Java library for mathematics, linear algebra and optimisation.
• exp4j is a small Java library for evaluation of mathematical expressions.

(c) Wikipedia