Maxwell-like equations are offered that are valid not only for electromagnetic, but also for gravitational fields. They are based on an expanded form of the law of energy conversion, which explicitly takes into account the nonequilibrium of systems and the ability of energy to transfer from one energy carrier to another. This makes it possible to eliminate the distinction between electromechanics and the field theory, to find out the meaning of the magnetic vector potential as a function of the charge rotation speed, to reveal the tensor nature of the magnetic field and the presence of a divergent component of the scalar nature in it. The existence of the Lorentz magnetic forces and the presence of a work-performing moment therein are proved. Maxwell-like equations do not contain field operators and have a simpler form covering nevertheless a broader spectrum of phenomena due to taking into account the convective components of bias currents. Wave equations alternative to the Maxwell ones are offered and the non-electromagnetic nature of light is substantiated.

RNN and strengthen the semantic information of the context, BLSTM is used to replace the RNN model for label prediction, and then the CTC algorithm is used to complete the transcription and output the final recognition result. Experimental results show that the improved CRNN text recognition algorithm has an accuracy rate of 96.6%, which is 1% higher than the basic CRNN text recognition algorithm, and this end-to-end network structure design also greatly shortens the text recognition time.

In recent years, housing prices across the country have been on a slow upward trend, and the real estate market is crisscrossed, and Shenyang is no exception. This paper chooses the Elman neural network that can process dynamic time series information to predict the trend of average housing prices in Shenyang. The article selects the average house price of Shenyang for 72 months from September 2015 to August 2021 in Shenyang, using 7 months as a set of training samples, the first 6 months as input data, and the seventh month as output data .There are 66 groups in total, the first 58 groups are used as training sets, and the last 8 groups are used as test sets. At the same time, the BP neural network and the RBF neural network are used for the same prediction, and the prediction results of the three are compared and analyzed. It was found that both the Elman neural network and the BP neural network are better than the RBF neural network in predicting performance. In terms of error details, since the Elman neural network can better process time series data, compared to the BP neural network, the predicted value of the Elman neural network is closer to the true value, and the predictability is better.