Search

In this work we propose a self improving artificial intelligence system to enhance the safety performance of reinforcement learning RL based autonomous driving AD agents using black box verification methods RL algorithms have become popular in AD applications in recent years However the performance of existing RL algorithms heavily depends on the diversity of training scenarios A lack of safety critical scenarios during the training phase could result in poor generalization performance in real world driving applications We propose a novel framework in which the weaknesses of the training set are explored through black box verification methods After discovering AD failure scenarios the RL agent s training is re initiated via transfer learning to improve the performance of previously unsafe scenarios Simulation results demonstrate that our approach efficiently discovers safety failures of action decisions in RL based adaptive cruise control ACC applications and significantly reduces the number of vehicle collisions through iterative applications of our method The source code is publicly available at https github com data and decision lab self improving RL

Myasthenia gravis (MG) is an autoimmune disease caused by antibodies targeting the neuromuscular junction of skeletal muscles. Triple-seronegative MG (tSN-MG, without detectable AChR, MuSK and LRP4 antibodies), which accounts for ~ 10% of MG patients, presents a serious gap in MG diagnosis and complicates differential diagnosis of similar disorders. Several AChR antibody positive patients (AChR-MG) also have antibodies against titin, usually detected by ELISA. We have developed a very sensitive radioimmunoprecipitation assay (RIPA) for titin antibodies, by which many previously negative samples were found positive, including several from tSN-MG patients. The validity of the RIPA results was confirmed by western blots. Using this RIPA we screened 667 MG sera from 13 countries; as expected, AChR-MG patients had the highest frequency of titin antibodies (40.9%), while MuSK-MG and LRP4-MG patients were positive in 14.6% and 16.4% respectively. Most importantly, 13.4% (50/372) of the tSN-MG patients were also titin antibody positive. None of the 121 healthy controls or the 90 myopathy patients, and only 3.6% (7/193) of other neurological disease patients were positive. We thus propose that the present titin antibody RIPA is a useful tool for serological MG diagnosis of tSN patients.

Seronegative myasthenia gravis (MG) presents a serious gap in MG diagnosis and understanding. We applied a cell based assay (CBA) for the detection of muscle specific kinase (MuSK) antibodies undetectable by radioimmunoassay. We tested 633 triple-seronegative MG patients' sera from 13 countries, detecting 13% as positive. MuSK antibodies were found, at significantly lower frequencies, in 1.9% of healthy controls and 5.1% of other neuroimmune disease patients, including multiple sclerosis and neuromyelitis optica. The clinical data of the newly diagnosed MuSK-MG patients are presented. 27% of ocular seronegative patients were MuSK antibody positive. Moreover, 23% had thymic hyperplasia suggesting that thymic abnormalities are more common than believed.