Study Finds Racial Bias In Facial Recognition Tools
A number of facial recognition tools available today have a difficult time when it comes to identifying Asian, African-American, as well as native groups as compared to Caucasians. This observation is based on a study by the National Institute of Standards and Technology or NIST.
False positives were higher among the groups mentioned previously when confirming if a photo is a match to another image of the individual in the database. This is a matching type is called one-to-one matching. It is usually utilized for unlocking phones and checking passports.
The study by the NIST looked into 189 software algorithms coming from 99 developers. The companies included were Panasonic, Intel, and Microsoft. According to the NIST, Amazon did not submit its software algorithm to be tested. The Recognition software of Amazon has received criticism about having racial and gender bias. So far, Amazon does not have a comment regarding this matter.
Women received more false positives compared to men. It is a consistent finding across the datasets and algorithms, the NIST stated. Among children and the elderly, there was also a higher rate of false positives.
Patrick Grother, a computer scientist of the NIST stated that he and his team have gathered empirical evidence across facial recognition algorithms for having demographic differentials. While they are not exploring the cause of the differentials, having the data can be valuable to developers, policymakers, as well as end-users when it comes to the appropriate use and also the limitations of the algorithms.
Another type of matching which is known as one-to-many matching is about determining if an individual in a photo matches another in the database. A person of interest can be identified using this type of matching. The study of the NIST found that this matching type has higher false positives among African-American females. It is vital to look into this because such false positives can lead to accusing the wrong person of an act which they did not commit.
For the one-to-one search, having a false negative can only result in a slight inconvenience which is you not being able to unlock your phone. The second attempt is usually successful. However, when it comes to the one-to-many search, having a false positive will put an incorrect match among candidates who will be under scrutiny.
What the study emphasized was that not all of the facial recognition algorithms have high rates of false positives in the one-to-many matching type. In the end, different facial recognition algorithms perform in different ways. Nevertheless, the gender and racial bias of the facial recognition tools are a cause for concern.
There have been a few previous studies that tried to look into the demographic effect of the one-to-one matching type. However, the demographic effect of the one-to-many matching type is yet to be explored, the NIST stated.
The NIST conducted this study using 4 collections of images. It has 18.27 million photos of 8.49 million individuals. These were taken from the databases that the State Department, FBI, and the Department of Homeland Security provided.
What Is REAL Messenger?
REAL Messenger is an app designed to help agents promote their listings and themselves.
Social media is table stakes now. But not all apps are created equal. Facebook is for friends. Instagram is for interests and LinkedIn is for work connections. So why shouldn’t real estate has its own social media platform that brings together a global community of agents, buyers, and sellers?
We’ve learned that there’s power in a platform to find what you’re looking for, to share, and to chat — opening doors and elevating your presence. We’ve watched how people become influencers with followers who devour every post. These influencers don’t have to pay to promote themselves, they’ve built their audience through content, personality, and what they stand for.
In contrast, real estate has become “pay-to-play,” restricting agents from showcasing their listings on well-known real estate platforms, unless they pay (a lot!) to promote them. Agents have lost control in the real estate process, while big proptech profits from agents’ hard-earned listings.
Social media meets real estate
Imagine a social media outlet just for real estate — one like Instagram or WhatsApp geared 100% to our industry. The audience is engaged in real estate. Agents connect to share information about listings with one another or potential buyers and sellers – and retain those connections. Agents can even share their knowledge of properties before they are listed publicly.
It’s now all possible with the REAL Messenger app, an incredibly fast social media platform for real estate agents to promote their listings and share their styles and specialties, as well as their sales history and approach. Integrated into the app is an easy chat feature that replaces the need for cold calls, excessive emails, and online ads that don’t yield much return on investment.
Giving agents back control
From the agents’ perspective, sites and apps like Zillow are taking listing information from MLS agreements, repackaging it to promote it on their sites, then selling the information back to the agents who owned it in the first place! Agents end up paying these sites expensive advertising fees. And while many agents use Instagram to let their followers know about their listings, they are not really targeting a real estate-specific audience. WhatsApp is also used for secure, data-encrypted conversations ensuring quick exchanges of information. But how can an agent build their business when they’re promoting to people who are not in the market? Our formidable team of developers created the best of all worlds for the world of real estate — it’s like Instagram with a secure chat feature similar to that of WhatsApp.
The REAL advantage
Self-branding and inbound marketing is built into REAL. Agents brand themselves by creating content that showcases their listings, providing information buyers will need, and sharing their successes with transactions. Agents use the app’s three-point rating system to describe their listing and other important characteristics.
Potential buyers can search for anything specific to their interests (i.e., a home with a patio, a garden, or a swimming pool) in particular zip codes. They can also browse by scrolling through the listings to find the hottest, most popular real estate properties in their areas. These potential buyers can follow agents whose posts resonate with their preferences and interests, expanding agents’ networks.
Deepfakes are fake videos created using digital software, machine learning, and face swapping. Deepfakes are computer-created artificial videos in which images are combined to create new footage that depicts events, statements, or actions that never actually happened. The results can be quite convincing. Deep fakes differ from other forms of false information by being very difficult to identify as false.
How do deepfakes work?
The basic concept behind the technology is facial recognition, users of Snapchat will be familiar with the face swap or filter functions which apply transformations or augment their facial features. Deep Fakes are similar but much more realistic. Fake videos can be created using a machine learning technique called a “generative adversarial network” or GAN. For example, a GAN can look at thousands of photos of Beyonce and produce a new image that approximates those photos without being an exact copy of any one of the photos. GAN can be used to generate new audio from existing audio, or new text from the existing text – it is a multi-use technology. The technology used to create Deep Fakes is programmed to map faces according to “landmark” points. These are features like the corners of your eyes and mouth, your nostrils, and the contour of your jawline.
When seeing is no longer believing
While the technology used to create deep fakes is a relatively new technology, it is advancing quickly and it is becoming more and more difficult to check if a video is real or not. Developments in these kinds of technologies have obvious social, moral, and political implications. There are already issues around news sources and the credibility of stories online, deep fakes have the potential to exacerbate the problem of false information online or disrupt and undermine the credibility of and trust in news, and information in general.
The real potential danger of false information and deep fake technology is creating mistrust or apathy in people about what we see or hear online. If everything could be fake does that mean that nothing is real anymore? For as long as we have had photographs and video and audio footage they have helped us learn about our past and shaped how we see and know things. Some people already question the facts around events that unquestionably happened, like the Holocaust, the moon landing, and 9/11, despite video proof. If deepfakes make people believe they can’t trust video, the problems of false information and conspiracy theories could get worse.
False news can lead to false memories
One of the most common concerns and potential dangers of deep fakes and false information, in general, is the impact it can have on democratic processes and elections.
A recent survey from UCC confirmed that people recall fake news more than real news. The results of the survey indicated that voters may form false memories after seeing fabricated news stories, especially if those stories align with their political beliefs, according to a new study. The researchers suggest the findings indicate how voters may be influenced in upcoming political contests, like the 2020 US presidential race.
The author of the report Dr. Gillian Murphy added; “This demonstrates the ease with which we can plant these entirely fabricated memories, despite this voter suspicion and even despite an explicit warning that they may have been shown fake news,”.
What Is Cognitive Computing?
Cognitive computing is the use of computerized models to simulate the human thought process in complex situations where the answers may be ambiguous and uncertain. The phrase is closely associated with IBM’s cognitive computer system, Watson.
Computers are faster than humans at processing and calculating, but they have yet to master some tasks, such as understanding natural language and recognizing objects in an image. Cognitive computing is an attempt to have computers mimic the way a human brain works.
To accomplish this, cognitive computing makes use of artificial intelligence (AI) and other underlying technologies, including the following:
- Expert systems
- Neural networks
- Machine learning
- Deep learning
- Natural language processing (NLP)
- Speech recognition
- Object recognition
Cognitive computing uses these processes in conjunction with self-learning algorithms, data analysis, and pattern recognition to teach computing systems. The learning technology can be used for speech recognition, sentiment analysis, risk assessments, face detection, and more. In addition, it is particularly useful in fields such as healthcare, banking, finance, and retail.
How Does Cognitive Computing Work?
Systems used in the cognitive sciences combine data from various sources while weighing context and conflicting evidence to suggest the best possible answers. To achieve this, cognitive systems include self-learning technologies that use data mining, pattern recognition, and NLP to mimic human intelligence.
Using computer systems to solve the types of problems that humans are typically tasked with requires vast amounts of structured and unstructured data fed to machine learning algorithms. Over time, cognitive systems are able to refine the way they identify patterns and the way they process data. They become capable of anticipating new problems and modeling possible solutions.
For example, by storing thousands of pictures of dogs in a database, an AI system can be taught how to identify pictures of dogs. The more data a system is exposed to, the more it is able to learn and the more accurate it becomes over time.
To achieve those capabilities, cognitive computing systems must have the following attributes:
- Adaptive. These systems must be flexible enough to learn as information changes and as goals evolve. They must digest dynamic data in real time and adjust as the data and environment change.
- Interactive. Human-computer interaction is a critical component of cognitive systems. Users must be able to interact with cognitive machines and define their needs as those needs change. The technologies must also be able to interact with other processors, devices, and cloud platforms.
- Iterative and stateful. Cognitive computing technologies can ask questions and pull in additional data to identify or clarify a problem. They must be stateful in that they keep information about similar situations that have previously occurred.
- Contextual. Understanding context is critical in thought processes. Cognitive systems must understand, identify and mine contextual data, such as syntax, time, location, domain, requirements, and a user’s profile, tasks, and goals. The systems may draw on multiple sources of information, including structured and unstructured data and visual, auditory, and sensor data.
Examples and applications of cognitive computing
Cognitive computing systems are typically used to accomplish tasks that require the parsing of large amounts of data. For example, in computer science, cognitive computing aids in big data analytics, identifying trends and patterns, understanding human language, and interacting with customers.
Featured3 years ago
Run Your Home The Smart Way: Apple HomeKit’s Best Products of 2020
Reviews3 years ago
The 10 Greatest & Irreplaceable iPad Apps
Featured3 years ago
Navigating The Soon-To-Release Honda E
Future3 years ago
Google CEO On AI: Regulation Of Artificial Intelligence Is Needed
Featured3 years ago
New Medical Bed Inspired By Star Trek Makes X-Rays More Affordable
Hardware3 years ago
Roku Streaming Stick Plus: Getting The Best Value For Your Money
Closure3 years ago
Bose Retail Will Be Wiped Out From North America, Europe, Japan, & Australia
Featured3 years ago
Galaxy S20: Samsung’s Latest Phone To Be Unveiled In February