What does Vectorspace AI offer, in layman’s terms?

Vectorspace AI’s algorithms look at data and find hidden similarities which can be applied to many types of businesses. We algorithmically generate millions of dataset permutations in real-time. These datasets are the ‘gasoline’ used to power unsupervised learning approaches in AI/ML. Specifically, they are used to:

  1. make money, a.k.a. generate or protect alpha on the long and short side with traditionals and cryptos. For instance, hedge funds would use it to monitor stock prices - if Vectorspace AI finds that when X stock price rises/falls, Y and Z also fall. The fund manager can then react before that happens to avoid losing money.

“Dexamethasone Announcement Could Have Made Hedge Funds A Fortune”

“Generating Alpha from Information Arbitrage in the Financial Markets with NLP Datasets: 水涨船高”

  1. increase the speed of innovation, novel scientific breakthroughs and enable scientific discoveries. Similarly, for a pharmaceutical company, Vectorspace AI could identify hidden and non-hidden relationships between the effects of drug compounds on certain conditions. In other words, Information that shows drug X, which is used to treat a condition, actually has benefited many patients with different conditions - or just as importantly, has a poor outcome for patients with certain conditions. In effect the “Smart Baskets” give companies a huge early advantage (known as information arbitrage or alpha).

“COVID-19 Drug Repurposing Datasets Now Available in Collaboration with Vectorspace AI, Amazon & Microsoft”–microsoft-301030507.html

COVID-19 Dataset Builder:

Vectorspace AI services more than just the financial vertical defined by cryptos and traditional market vehicles. Our platform has applications in almost all industry verticals.

Why is the code not open source?

Github is private to protect algorithms and code. We have been software engineers since we have been kids in the 80’s. We are not going to allow other organizations pretend they are auditing our code while also paying them tons of ETH. That is one reason.

From our CEO: “I have dealt with my fair share of outside auditors while running a public company. I have also been through a few major software audits run by security firms (which we have done work for in the past) and standard software auditing done by firms like PwC. Auditing software is completely different from auditing your books.”

What are Smart Baskets?

Smart Baskets are the product of indexed entities such as stocks, cryptos, drug compounds, etc. that are grouped by a similar theme, enabling thematic investing, i.e., the ability to invest in prominent trends or themes. In doing so, we are able to extrapolate hidden relationships based on proprietary Natural Language Processing and Understanding (NLP/NLU) datasets. Baskets of cryptocurrencies or stocks are algorithmically generated based on a news event, topic or ‘special situation’ often used by internal market researchers inside hedge funds/asset management companies.

How are Smart Baskets generated?

Smart Baskets are generated based on an event, global trend or concept using our real-time NLU correlation matrix datasets. They are smart because they are automatically generated based on NLU datasets. For instance, our clients choose a customized dataset with a custom sector of equities and a custom set of topics, trends, categories, pre-select events, or real-time events. A basket is generated and then they choose how they want to filter that basket which can be done in a million different ways.

Will the Smart Baskets be available to the public as well as institutions?

Yes, these will be available for traditional retail traders as well i.e., crypto traders. The larger customers will typically access more advanced versions of these Smart Baskets based on the tiered pricing.

Smart Baskets is a known term within the trading community. Would it be a good marketing move to rename Vectorspace AI’s Smart Baskets to a term with exclusivity to Vectorspace AI?

This makes sense, but based on our experience with branding, there are advantages and disadvantages to branding collisions. We rebranded in the past from Starmine. This is a common practice.

I read that the Smart Baskets will be exclusively developed for LCX. Is this a good strategy?

Smart Basket technology comes in many different forms which are built on datasets. These datasets have billions of permutations, so if a class of datasets supports a group of Smart Baskets, then the team can make that exclusive. We are in the process of negotiating a variety of exclusive licenses related to aspects of our technology.

Does Vectorspace AI test the Smart Baskets with a high variance in position size?

Yes, we have done this with a group before. We had a chance to test baskets with a few million dollars in capital per basket. One of our goals is to support deploying a large amount of capital.

What is the upper limit of position sizes that Vectorspace AI tested/back-tested?

We have tested in simulated markets with 1 million dollars per basket. You can back test and test in a simulated market environment with buy and sell pressure but this is nothing like the real thing. The real markets are slightly different which is why we plan to trade our internal proprietary baskets ourselves.

In an unfiltered basket do the stock ticker symbols come up in any specific order, i.e., from best option to worst or vice versa? In other words, would the best one shows up 1st in line?

Yes, they are ranked/scored based on the strength of the relationship. They are unfiltered because the relationships are not context controlled or directional.

Long baskets were discarded due to COVID-19 making it not viable, but we are currently seeing retail investors making a killing with Robinhood buying bankrupt companies. Would this new environment make long baskets not far more attractive than short baskets?

The first reason is because anytime you see someone making money based off of chapter 11 plays (bankruptcies), this is a chance play. These are not worth playing as it is a gamble. When you look at long baskets, they are not something we are focused on at this time. We started generating long Smart Baskets for the crypto markets which crashed. We then moved into datasets for every industry. Then we had an economic crash which resulted in a shift to short plays. The investment club will offer baskets for shorts on the traditionals side. Baskets are certainly available, and in this environment, we see them as being a big opportunity which we are prepared for. Institutions and retail will be offered to the opportunity to short traditionals.

The Coronavirus basket (COVID-19) resulted in a multitude of long opportunities. The secret sauce to Smart Baskets is the ability to take a Smart Basket and filter them by criteria of context to create two groups that allow you to go short and long. Creating that separation or filters is the goal. Data is the unrefined crude oil, while datasets are the refined petroleum that power AI systems. The dataset augmentation part is important because it allows data engineering pipelines to take their datasets and append them with Vectorspace AI NLP datasets. This will allow them to create new clusters/correlations as to why things are occurring.

Is there a way to test the Smart Baskets generation?

Yes, there is a demo version where you are able to set the themes and the algorithm will generate static thematic smart baskets containing the symbols of the public trading vehicles (stocks) related (containing known and hidden relationships) to each one of themes. The tool also includes a comparison of the smart baskets returns against the S&P index. You can try it at

These are demos designed to exemplify our capabilities. This is the tip of the iceberg. Demos are designed for NLP engineers inside hedge funds. Full commercial grade versions are only available to institutional customers.