Using artificial intelligence to predict "when a girlfriend gets angry"

All say the woman heart undersea needle, every guy who has been in love, may have been driven crazy by his girlfriend's words. You may never know when and why she is angry, then there is no black technology, can predict when the girlfriend will be angry?

The magic of artificial intelligence can be.

Now, artificial intelligence forecasts have accumulated a number of successful cases, able to analyze video streams and identify thousands of topics, and use AI techniques to analyze content in depth. In the use of artificial intelligence to predict "girlfriend when angry" issue, involving basic, technical, application of several aspects.

Base Layer

For the prediction of emotions, involving a large amount of data, call and artificial intelligence to match the new cloud computing architecture, data storage, tuning to achieve millisecond transmission, otherwise everything will appear meaningless; To do this, we first need to collect data about your girlfriend's anger and delight: basic dimensions, intrinsic factors, external factors, collateral factors, Large data such as potential factors, duration, angry consequences, mitigation time, mitigating causes, and easing results. Then, there are countless small sub-items, for example, from the expression to the action, from the facial features to determine the change, the head and limbs will produce what actions, words will say what words, to record, to extract the words of high frequency words. The above items are continuously refined into small items, a large number of records, and finally the use of the table to classify the phenomenon, the data, it is best to reach the level of big data.

Also avoid the "predictive as intervention" trap. For example, when we predict "when your girlfriend is angry," the focus is on relevance, which is a logical and common method of big data analysis, but it also needs to be as clear as possible about the link behind it, but the correlation between the label data and the anger is obviously very complex, The precise correlation between every data and anger is obviously another big job. For example, your girlfriend is very happy to be praised by the leader in the unit. However, to study why the praise, mining direct factors, comprehensive factors is a more troublesome thing, from a sociological perspective, may be a more abrupt or unreasonable things. Therefore, we have to ask for the next-fine-tuning algorithm, so that the machine repair before the test is not allowed, but each repair caused another error. Like the "Uncertainty principle" in the Heisenberg (Werner Heisenberg) Quantum world, there is a "predictive interference" trap in the big data world. How to avoid this trap depends on the increase in computational power or the advent of new algorithms.

Note: Be sure to get your girlfriend's consent. Imagine, if your girlfriend does not know, you secretly study her, which day unmasked, the consequences may be disastrous, although your starting point is good, but it is not important ....

In addition, data structure transformation, the credibility of the information, the data aggregation model, the automatic labeling of data, the data set can be trained, high sensitivity, high-precision sensor research and development, is also a very key issue, or "When your girlfriend angry" The emergence of countermeasures is not worth;

Therefore, cloud computing, big data, and the systematic coordinated matching of the internet of Things are the basic functions of predictive intelligent hardware.

Technical Layer

Chips and new algorithms are always the eternal theme. At present, the algorithm technology is heuristic. Whether the algorithm will perfectly solve a given problem is unclear, there is no mathematical theory to indicate whether a "good enough" algorithm solution exists. The algorithm is heuristic, and the work represents the effective.

This requires artificial intelligence to have enough learning depth and speed. For example, a two-year-old can be told several times after the elephant recognition, AI need to collect thousands of "girlfriend angry" samples, to accurately judge, your girlfriend will be angry.

Application Layer

Compared to predicting when a girlfriend will be angry, more importantly, "How to prevent a worse situation in time", Ai in the case of large numbers of data, can analyze "girlfriend will be angry", thus giving "how to effectively eliminate" measures.

--------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------