Can AI be Rescued, USA Covered by Shot Haze

图片描述

On August 3 and 4, local time, two mass shootings occurred in the United States, killing 31 people and injuring 51. In the United States, where shootings are almost daily, many companies use artificial intelligence technology to locate the shooting site in real time, predict criminal behavior, and successfully reduce the crime rate.

In the past weekend, two US states (Texas and Ohio) had a series of gunfire. Locals were horrified by these two nightmarish shootings.

图片描述

  • Flowers were placed at the scene of the El Paso shooting to pay tribute to the dead.

The mass shooting in El Paso, Texas, killed 22 people and injured 24, making it the seventh deadliest shooting in US history.

Second Bill of Rights: The Haze of Guns Over 200 Years in America

On the morning of August 3 (Saturday) local time, a shooting occurred in a Wal-Mart hypermarket in El Paso, Texas. A 21-year-old white man armed with AK-47 strafed the residents present, killing 22 innocent people.

Less than 14 hours after the Texas shooting, when everyone was still in shock, Dayton, Ohio, was immediately followed by a major shooting in the early morning of the 4th local time, killing 9 people and injuring 27.

Shooting has become the most common cause of death in the United States. According to the statistics of the non-profit Gun Violence Archive, 253 mass shootings have taken place in the United States so far this year (the latest data at the time of publication).

图片描述

  • The death toll from shootings in the United States in 2019 has reached more than 8,400

Two consecutive shootings within 24 hours have left Americans in anxiety and fear. They held a demonstration demanding that the US government and Congress actively push forward the gun control law.

However, the legal history of legal gun ownership in the United States has been a long time, forming a deep-rooted stubborn disease of gun violence, which has enveloped the United States for more than 200 years like a haze.

On December 15, 1791, the Second Bill of Rights (also known as the Second Amendment to the U.S. Constitution) was approved, namely:

A well regulated militia, being necessary to the security of a free state, the right of the people to keep and bear arms, shall not be infringed.

“A well-trained militia is really necessary for the safety of a free state, and the right of citizens to own and bear guns is inviolable. 」

图片描述

  • Domestic netizen comments: Free America, Shooting Every Day

Philip J. Cook, an American criminologist, speculated that if guns were not so widespread, criminals might still commit crimes, but at least would use other less lethal weapons.

AI Anti-terrorism and Anti-violence: Police’s Winning Weapon

After the shooting, US netizens posted that “it is not safe anywhere” and others were worried that they needed an escape plan at any time in public places.

If one hopes for the enactment of the U.S. gun control law, there will be too much uncertainty, and no one knows when this day will come.

Therefore, U.S. government agencies and many companies are actively trying to use AI to reduce the occurrence of crimes from the perspective of crime prevention and early warning.

Similar to the robot “prophet” in the movie minority report, it can detect people’s criminal attempts, so before someone commits a crime, the crime prevention organization will arrest him to avoid causing crimes and injuries.

Alarm Sharp Weapon: Lock the Shooting Site Immediately

Three masters and doctors from Massachusetts Institute of Technology, Stanford University and Rice University jointly founded ShotSpotter. The company uses sensors, artificial intelligence algorithms and systems to locate the shooting scene in real time through surveillance of public places to help law enforcement departments curb gun violence.

图片描述

  • Schematic of the working principle of ShotSpotter

They detected the fire of the gun and divided it into the following three steps:

Step 1: Real-time Monitoring and Location

Place acoustic sensors in public places and set software to filter out environmental background noise, such as car horns. When gunfire was heard, the sensor detected a pulse. After at least three sensors detect the pulse of the gunshot, the time difference between the sound received by the sound multi-point positioning sensor and the sound analyzed by the machine learning algorithm are used to determine the shooting position. The information is then sent to the Incident Review Center.

Step 2: Acoustic Evaluation and Classification

The machine classifier compares the sound with a large pulse database of known shots and other community sounds to determine if it is a shot. The IRC’s acoustic experts conducted further analysis and added relevant tactical intelligence, such as “multiple shooters” and “automatic weapons.” This may change the way the police approach the crime scene.

图片描述

  • Judging Gunshots by Sound Pulse

Step 3: Alert Notification

After the event is confirmed, the event notification will be sent to the dispatch center, MDT of patrol car, smart phone, desktop, browser and other applications. From the initial detection of gunfire to the alarm, the total time is less than 60 seconds. In addition, the alarm can integrate and trigger other systems (such as cameras) to pan, tilt and zoom in the shooting area to understand the real-time situation.

According to ShotSpotter, their system can provide the authorities with real-time and effective information about the type of shooting, and the location can be as accurate as 10 feet.

At present, their equipment has been deployed in 100 cities such as new york, Chicago and Santiago. Most of their customers are in the United States. Cape Town in South Africa also became their customers last year. Last year, under the deterrent of this equipment, the shooting rate in Cincinnati dropped by nearly 50%.

A sharp weapon for prevention: predicting high crime areas.

Predpol is a company that uses big data and machine learning to predict when and where crimes will occur. They claim that by analyzing the existing data of past crimes, the time and place where new crimes are most likely to occur can be predicted.

At present, their systems are being put into use in several cities in the United States, including Los Angeles.

图片描述

  • Predpol predicts the possible future crime location (right) through the heat map of the common crime location (left)

The team worked with mathematicians and behavioral scientists from UCLA and Santa Clara University to evaluate various data types, behaviors and prediction models. These models were further refined by crime analysts and officials from the Los Angeles Police Department and Santa Cruz (California) Police Department.

They finally determined the three most objective data points collected by the police department to provide the most accurate input data for the prediction: crime type, crime location, crime date and time.

Their algorithm is based on the observation that certain types of crimes tend to aggregate in time and space. For example, if a house is stolen today, the risk of being damaged tomorrow will actually increase, while the neighbors of the house will also face greater risks.

They call this technique real-time popular aftershock sequence crime prediction. Their system highlights possible crime sites on the map, and the police should consider more patrols.

图片描述

  • Chairman Chris Kelly holds a master’s degree from Yale University and a doctor’s degree in law from Harvard University.

Combating Crime, Global Consensus

It is estimated that the total expenditure on law enforcement in the United States exceeds 100 billion US dollars per year, which is only a small part of the government’s expenditure on crime. In addition, it also needs to bear some medical expenses and public property losses caused by crime.

Therefore, it is urgent to reduce the crime rate in the country and among the people. Artificial intelligence technology also shows its advantages of rapidity and accuracy in reducing crime rate.

Existing systems that use artificial intelligence and machine learning to detect crimes through sound or video cameras have proved effective and are expected to continue to expand.

图片描述

  • After all, Zhang Xueyou has captured fugitives through face recognition in many concerts.

We believe that the development of technology will definitely make the world a better place.