Got to know Java Virtual Machine can be used in FPGA's. Besides, a Java Optimized Processor has also been developed. Will this help us implement Machine Learning as Hardware Implementation using logic gates in near future?
Maybe this is a good starting point: http://www.jhdl.org/ . Alternatively you may check of hardware implementations of the JVM. Martin Schoeberl has done JOP: http://jopdesign.com/ and there are possibly some other implementations as well, like PicoJava.
I'm a big fan of JHDL and Java. But I do not understand the question. What implication is there that having Java in FPGAs can benefit machine learning? Machine learning is already a very abstract term. Could you ellaborate a little bit more the question?