R2R Analytics works with publicly traded companies and provides stock market trade surveillance, social media analysis, investigation, and data analysis services to provide a 360° view of the forces influencing a company's share price and corporate reputation. We help companies anticipate challenges and identify opportunities and make better, more informed decisions.Social media may be weaponized to influence investor sentiment, and corporate reputations and manipulate trading algorithms.
As recent events have shown, social media
The influence of social media on the stock market is undeniable is a primary vector through which a company’s share price may be influenced.
On January 28, 2021, U.S.-based video game retailer GameStop Corp. saw its stock reach an all-time high of $483. Two weeks earlier, the stock had been trading at $20. Two weeks later, it was down again, and a congressional hearing on the matter was underway in Washington. The GameStop case is just the latest example of how social media is being harnessed by groups of individuals to coordinate behavior and move the share prices of publicly traded companies.
Flash crashes
On 04/23/2013 hacker a AssociatedPress’s Twitter account Tweeted the fake breaking news headline “Two Explosions in the White House and Barack Obama is Injured”) erased $136 billion market value in just 60 seconds. Flash crashes
Collusion, organizing, spreading disinformation, creating echo chamber of negativity.
70% of total market trading volume controlled by algorithms, many are designed to automatically trade on breaking news, sentiment in milliseconds. Arms race to take advantaged of information asymetries measured in milliseconds.
Traditional approaches to social media analysis are ineffective.
distinguish between retail and professional
traditional metrics of determining influence don’t apply
Social graph analysis to identify connections and associations
Manipulating stock prices with an adversarial tweet
Researchers show that stock price prediction models can be tricked by simply quote-tweeting an influencer’s post on Twitter and altering a word or two.
Some of the world’s biggest investors are turning to machine-learning models to guide their trading decisions. These models scan the web for clues that stocks are under or overvalued and should be bought or sold. But a new study1 suggests that these predictive models may not be as trustworthy as they might seem. The study, which was presented at the natural language processing conference NAACL, found that these models are often biased against certain companies and industries. This bias can lead to inaccurate predictions about stock prices, which could spook investors and cause them to make bad decisions.
Using publicly available data and prediction models, the researchers built a tool that can identify from a series of influencer tweets the one that appears easiest to attack. The tool then finds the word in the target tweet that’s most likely to flip the model and swaps it with a semantically similar word when it quote tweets the original. The substitute word is unlikely to raise any red flags among readers because of its similar meaning, but it causes the model to reverse its prediction.
After ingesting a doctored tweet, a model that might have predicted that a stock price was falling and suggested that investors sell, might reverse its decision, and instead nudge the investor to buy. “The fake retweet can fool the stock prediction model, but the human eye is unlikely to notice that anything has changed,” said IBM researcher and senior author of the study, Pin-Yu Chen.
“Machine learning brings new risks to financial decisions,” said Xie. “With pensions and college savings at stake, we need to understand where the vulnerabilities are and how to reduce them.”
Original tweets can’t be edited, but you can say whatever you want in a quote tweet. It’s this vulnerability that Wang and his colleagues exploit in a new technique for launching adversarial attacks on deep learning-based stock-prediction models. Their experiments may be the first to probe the weaknesses of financial models that base their forecasts, in part, on news gathered from social media.
“Our results show that the proposed attack method can achieve consistent success rates and cause significant monetary loss in trading simulation by simply concatenating a perturbed but semantically similar tweet.”
“If you want to manipulate stock prices, you don’t need access to an investor’s model or data,” said Dakuo Wang, a researcher at IBM. “You just create a few hundred fake Twitter accounts, pretend to be an investor, and change a word or two when quote tweeting the CEO.”
For example,Although the event doesn’t fall into the category of adversarial attack, it rings the alarm for traders who use (social) media information for their trading decisions.
Attack model: Adversarial tweets. In the case of Twitter, adversaries can post malicious tweets which are crafted to manipulate downstream models that take them as input.
Secondly, adversarial tweets are optimized to be semantically similar to the original tweets so that they are not counterfactual and very likely to fool human sanity checks as well as the Twitter’s content moderation system.
11 May 2022 research published on July 2022
Figure 1: An example of word-replacement adversarial attack. (Top) benign tweet leads Stocknet to predict stock going UP; (Bottom) adding an adversarial quote tweet leads Stocknet to predict stock going DOWN. It is now known that text-based deep learning models can be vulnerable to adversarial attacks
Since 2013 Bloomberg has provided Twitter analysis feed for institutional clients to develop trading algorithms.
One of the key challenges on social media is evaluating content to discern the
Social media has democratized publishing but become one of the most powerful tools in the stock market manipulator's arsenal.
In the past, stock market manipulation was a tedious and expensive process.
Identify currents, themes and detect inauthentic content designed to influence public perception, investor sentiment and manipulate markets.
Not all social media chatter is the same. For analysis to have meaning, it’s critical to be able to differentiate between legitimate, frustrated shareholders and short sellers engaged in a bashing campaign. Threat analysis retail versus a sophisticated investor or financial industry pro.
It's also critical to be able to prioritize channels and differentiate between
Detect and respond to social media
Social media is a vast but nebulus
Traditionally, social media monitoring was relegated to the investor relations department but
Graph AI is a fully integrated next generation technology, leveraging artificial intelligence, machine learning and natural language processing (NLP) to extract meaning from text and provide deep contextual understanding of social media content.
False rumors spread faster and wider than true information, according to a 2018 study published in Science by professors from MIT Sloan Business school and MIT Media Lab. They found falsehoods are 70% more likely to be retweeted on Twitter than the truth, and reach their first 1,500 people six times faster.
The novelty hypothesis, which found that people are drawn to information that is novel and unusual, as false news often is.
Fully integrated analysis
Bots
Disinformation / misinformation detection
manipulate investor sentiment
Manipulate trading algorithms
https://www.businessinsider.com/fake-muddy-waters-tweet-hoax-2013-1
Social media campaigns designed to damage brands and spread disinformation. Starbucks & Nike fake coupons
Meme Stock Fallout - retail investors are more insistent that publicly traded companies monitor and respond to negative social media.
False short reports
According to MIT Sloan researchers, falsehoods are 70% more likely to be retweeted on Twitter than the truth, and reach their first 1,500 people six times faster.
KEY DIFFERENTIATION
Social intelligence and social listening tools are helpful in flagging activity and volume, but they don’t provide insight into the source of a narrative, influencers engaged in the conversation, the impact on key audiences, or where the conversation is heading.
Other social media monitoring solutions are ineffective at detecting disinformation and influence operations and are incapable of linking online behavior to changes in a company’s share price. Nor can they anticipate attacks and identify those responsible.
Social media driving sentiment & algos
Over 75% of investors rely on social media to inform their investment decisions and an increasingly large number of hedge funds rely on social media analysis to drive trading algorithms.
In periods of market uncertainty, investors become more risk-averse and susceptible to selling based on unfounded allegations and rumors.
Get unparalleled influence detection with advanced AI & expert social media analysis.
Social media analysis is too important to rely on generic “social listening” applications.
Social media Is driving trading algorithms & investor sentiment
Social media has become ground zero for stock market influence and manipulation. R2R Analytics fills painful intelligence gaps to ensure companies are the first to know when influence and information are spreading online.
Disinformation analysis pioneers in disinformation analysis
Features Overview
Identify
Traditional social media analyses are not designed to analyze stock market chatter, identify how social media is influencing share prices, flag which accounts are attempting to influence share prices or save critical data that could be evidence in a future investigation.
See how R2R 360° beats traditional social media monitoring.
R2R 360° offers the most sophisticated social media analysis available. Experienced analysts leverage natural language processing, and proprietary machine learning algorithms to actively track how information is traveling across social media, flag suspicious activity, identify the accounts influencing investor sentiment and share price and provide stakeholders with detailed analysis and reporting.