Mircea Dan Hernest
About
I am currently invested into multi-class image classification based on the binary classification made public by Google Research (so-called AutoML-Zero). Fresh experience with C/C++ (including OpenMP) and Python (https://github.com/rist-ro/google-research/tree/master/automl_zero). Previous experience with Cloudera: Hadoop, Spark, Scala, Impala, Kudu, Kafka, Jasper Reports. Older experience with Java, Castor, JAXB, Spring, Hibernate, JAX-RS, Maven, Eclipse. Generally interested in certified program synthesis. I have been a researcher specialized in certified and automatized program synthesis from data. In my research I co-developed "light functional interpretations" which are direct adjustments of Goedel's `Dialectica' interpretation for functional program extraction (from sesqui-intuitionistic proofs of input-output correspondence) without the detour to Linear Logic. http://hernest.eu/
Publications
-
Selected publications
-
2021
-
2010
-
2009
-
2008
-
2007
-
2005
-
2003