Fair Play, found within a complex blend of elements
The term 'bias' can mean different things in different domains. Statistical bias is when an operation is disproportionately weighted to favour some outcome. Social bias is when such operations relate to people, which may lead to unfair decisions being made.
Reducing bias is very challenging due to the complexity of data and models, as well as potential differing views on whether something is socially biased even if it may be statistically accurate. For example, men as a group are physically stronger than women as a group. However, we typically consider gender to be a protected characteristic, which is not permitted to unduly influence decisions in hiring, etc. This means that a system may be technically correct, yet still problematic with regard to the law.
Even data put through a high-pass filter in order to try to obfuscate inaccurate machine perceptions, to a degree that a human could never recognise it as an x-ray, may still contain signatures that machine learning can recognise.
Bias can sneak in from a number of sources, for example:
1). The reproduction of human labelling or selection biases, such as an algorithm that is trained upon human appraisals of resumes may replicate the same biased patterns.
2). Bias due to error in datasets, for example incorrect geolocation data that wrongly states that a house is inside a lake, and therefore considered not able to insured.
3). Bias due to a lack of sampling data, for example an algorithm that is trained upon a set of examples over-representative of one ethnicity or gender, which generalizes poorly to underrepresented demographics in the real world.
4). Bias due to overfitting, whereby a model is trained too strongly on training data, to the degree that it maps poorly onto real world examples.
5). Bias due to adversarial error, whereby a model may fail to recognise something accurately, or may misinterpret one thing for another. Models can be reverse engineered to uncover such exploits.
We can take several steps to reduce the risk of bias within algorithmic systems.
1). Select data which appears to be minimally influenced by human perception or prejudice. This is challenging, as data generally needs to be labelled and annotated in order to be interpreted by machine intelligence.
2). Make datasets more inclusive. Ensure that data is gathered from as broad a sampling as possible, and indeed solicit less common examples to ensure that the data is more representative of a global population and global environments.
3). Ensure the accuracy and integrity of data as far as is possible. Perform tests to ensure 'sanity checks' upon data, to search for signatures of error, and to attempt to locate lacunae (missing data), and either repair it, or ideally set it aside. This kind of work is a core duty of data science, and much of these rather dull efforts are performed by legions of workers in less-developed nations for very small sums of money, with uncertain credentials.
4). Rigorously test models against real world examples. Often, a portion of training data is set aside in order to validate that the model is learning correctly. However, much as a battle plan only lasts until the first engagement with the enemy, lab results are not trustworthy. Systems must be tested live, in as broad a range of environments and demographics as possible in order to be validated as truly accurate and effective.
5.) Harden systems against attack and exploitation. Resources should be ring fenced in order to provide bounties for Red Teams to attempt to disrupt the algorithmic system. This can help to uncover issues long before they may occur 'in the wild' where real people may be affected.
Machine learning systems are increasingly enmeshed with our personal and professional lives. We interact with algorithms a hundred times a day, usually without even realising. It's crucial that such technologies are not applied to exclude anyone, or allowed to unfairly misinterpret people's behaviour or preferences.
It's crucial that we embed transparency within algorithmic systems, so that we can understand what processes are being performed, in what manner, for what purposes, and to whose benefit. This can help to provide insights regarding biases within such systems also.
AI has tremendous potential benefit within our society, but there are strong risks of it turning into a prejudiced petty tyrant also. More governmental, academic, and business resources must be devoted to ensuring that we integrate AI safely and securely into our global society.
Natural Language processing is the science of teaching computers to make sense of the kinds of language that human beings use in everyday life. It is one of the most mature forms of machine learning, and highly integrated in daily life. NLP techniques assist speech processing, by helping to provide context for speech recognition systems. For example, the spoken word bear might mean an animal, or it might mean to carry or endure. If spoken, it could actually be intended as 'bare' as in nude, or a name, as in Behr. However NLP can guide the interpretation of such phonemes, to parse common phrases, and to make an inference that the word bear next to arms most likely refers to brandishing weapons, instead of furry ursine appendages.
However, not all cultures or subcultures use language in the same way. Words and expressions can mean different things. In the US one might pay the check with a bill, and in the UK one might traditionally pay the bill with a cheque. It's important that NLP attempts to understand not only the context of the surrounding language, but that it attempts to ascertain the probable culture in which someone is communicating. Furthermore, sometimes people may code shift, from a family dialect towards a more general way of communicating which others are more likely to find intelligible.
Another area of potential bias in speech recognition occurs with accents. The stress, tone, and pronunciation of a culturally idiosyncratic form of language may be ignored or misinterpreted by datasets which are trained upon a particular form of a language. This can lead to outcomes that are unfavorable for speakers of less recognised accents. In a world where algorithms make impressions about us in so many ways, such misinterpretation, or outright failure to understand, can lead to us being ranked less favorably for various opportunities, and may have a direct economic impact. It's important that we try to include a diverse range of examples from many different kinds of accents and dialects to help ensure that people are not unfairly excluded.
It's important that there is a clear mechanism to report failure or anomalies, as well as an opportunity to provide a corrective example, so that machine learning systems can improve over time, whilst performing their activities. It's also important to provide transparency as to what impression an algorithmic system made of a person, to ensure accuracy, appropriateness, and proportionality.