Interview: Vivien Lin, ​​Chief Product Officer at BingX

BingX is a leading cryptocurrency exchange, serving over 10 million users worldwide. We got to speak with BingX CPO Vivien Lin about how AI and blockchain are changing the way people trade online.

Blockchain is so much more than the technology that underpins cryptocurrencies. Vivien Lin explains how smart contracts and the decentralized nature of blockchain complement the analytical capability of AI.

[Q]: How can AI complement cryptocurrencies and how is it changing the way people are trading it?

[Vivien Lin]: “I think there are three different concepts. Cryptocurrency, blockchain technology, and AI technology, right? Cryptocurrency is more of a certificate. It’s something you hold because you want to invest in the value of the project.

You believe the value of the project is going to rise and you can benefit from that. But AI and blockchain are just technologies. 

If you’re trading a stock you can use AI to do what it is good at, which is to analyze data, to gather together information, like the price, the macro economies, micro-economies, everything, like data processing and it helps you see the trend of the market and to do the valuation, etc. 

When we’re talking about trading, I think there’s not much difference. The only difference is you’re trading a stock versus you’re trading a cryptocurrency.

Six or seven years ago, people trading in stocks or Forex, already deployed machine learning models to predict the trend and to make quantitative strategies. 

But now the methodology has obviously leveled up with AI, because the AI or the large language models we are now talking about, they have a much greater computation power. 

So it’s more efficient for us or for the trader who trades cryptocurrency or whatever asset to use a large language model or the latest AI model to make the prediction or to use them to auto-adjust the weights of the factor.

So this is how AI is implemented in investment and also cryptocurrency investment. One application is in, for example, stock trading or Forex trading. Much of the trading data is public. It’s centralized, but you find it’s very easy to get the data because every exchange publishes the trading volume, etc.

But in crypto, some of the transactions happen in a centralized exchange like ours. Then we will publish the volume or the price data. But a lot more transactions happen in a decentralized place.

So you have to have a tool to track all that data. And sometimes if there is a transaction across different blockchains or mainnet, you’ll find that for a human, it’s actually quite difficult for average people to track all this data. So if we can use AI technology, then the data tracking and data analysis will be much easier.

This is one of the implementations. And of course, as we operate our exchange, then we can use AI technology in the security space to detect suspicious activities, like in the anti-money laundry analysis, those kinds of tasks are currently where AI was used.”

[Q]: There’s a big concern about fraud when it comes to crypto trading and security measures. How is AI being integrated into these exchanges to ensure those security measures?

[Vivien Lin]: “Actually, I have to say this implementation is quite on early stage. Because if we’re talking about anti-money laundry or anti-cyber attacks, these are quite mature technologies. Some of them already integrate AI, and some part of them haven’t yet, but it’s quite mature technology.

Somewhere I think is more promising is to use AI technology to detect fraud, especially trading fraud. 

In traditional finance, because it’s highly regulated, you only have two stock exchanges in the US and one stock exchange in the UK. So those centralized exchanges are strictly regulated by the regulator.

It’s difficult for anyone trading on those exchanges to make a mistake or commit fraud. But if you’re looking at crypto, there are more than 200 centralized exchanges and I think over 1,000 decentralized exchanges. 

So it’s almost impossible at this moment for any regulator to regulate all of the exchanges.

So whether they can function properly largely relies on two things. One is how those decentralized exchanges or centralized exchanges regulate themselves. If they have set up a higher moral standard or higher ethical standard is one thing.

Another thing is to rely on their ability to detect fraudulent transactions. This could vary a lot. I would say if you’re looking at the top 20 centralized exchanges, I think they are ethically very good.

They don’t really want to make mistakes or they want to ensure that their business can last forever. But the thing is, do they have the necessary technology or necessary knowledge to support them in making the business last forever?

So before a company deploys AI technology it is highly reliant on the risk manager, if his or her knowledge is sufficient to write down all those cases, how people commit fraud or take advantage of the flaw of the trading rules in the exchange. 

But once the industry deploys more AI or more well-trained models, then even if the person who is in charge of risk management has a flaw in their methodology or flaw in the mechanism, AI will use the massive data to help us refine the design of the system.

So I think this is where AI is most helpful in emerging industries like crypto trading, where everybody is trying to gain more experience during the process. Sometimes people make mistakes. AI technology helps people reduce the chance of making mistakes.”

[Q]: How is AI changing using trading bots and copy trading systems? How is that changing the way that users are now shifting away from traditional methods of trading?

[Vivien Lin]: “If we’re looking at where the industry is at this moment, actually if you’re looking at the trading bots, they are quite simple. Just a great bot, right? But there are a few more advanced communities who have more experienced traders that want to start to implement AI-driven strategies.

It’s almost impossible for those crypto native traders to have the level of experience and the level of understanding of the traders in the traditional financial markets. So if you ask them to gather those more than 1,000 factors [indicators], it’s impossible for them.

But now they use AI to screen those factors to auto-adjust the weights of the factors the technology empowers that group of people to be able to make a strategy that is almost on par with those who come from the professional trading space.

Another thing is copy trading. In the past, a copy trader or master trader, they are human. So humans make mistakes, right? 

When you make a profit, you are reluctant to take profit. You always think that the price of the token could be higher and higher, right? But if you’re making a loss, you want to stay there. You don’t want to act on a stop loss.

So there’s always bias or human flaw in investment. But now with an AI strategy, it’s become easier for them to make a take profit or a stop loss decision. Or sometimes they are not aware, but their model tells them, it’s time to act on your stop loss or it’s time to take profit. 

I would say they use AI tools to assist them to analyze the market and to make a framework for them so that they have more confidence to follow the framework because they think, okay, maybe this is a summary of all those traders on the internet. So they have lower psychological hurdle to implement the rules strictly.”

[Q] What role is AI playing in helping traders refine their trading strategies when the consider market indicators or factors?

[Vivien Lin]: “I think it’s just about market prediction. In the past, if you were not a professional trader or you start to form your own trading philosophy but not there yet, at that time, you may be looking at tens, or a dozen, or several dozen factors and you start to feel that it’s hard to follow, right? 

Because some of the factors tell you to buy and some of the factors tell you to sell.

You don’t know how to read or how to translate all those factors. And now I think the best area for AI to step into the decision is it will help you to screen out those factors or those indicators that are not suitable in current market.”

[Q]: With data management, how does AI help to categorize and analyze this massive amount of data?

[Vivien Lin]: “Data management for me has two layers. One is how people like traders use AI to manage the data. It just boils down to what AI is best at, to summarize data and make the trend prediction and to screen out layers, those kind of things.

Another layer of the data management is in blockchain or in cryptocurrency. If we’re talking about blockchain technology instead of trading cryptocurrency, then some of the most promising sectors are such as DePIN. 

DePIN is like decentralized data management. One of the DePIN sectors is a decentralized data management system. It’s like a protocol which will sign the agreement with individual participants.

It could be a company or could be an individual. The protocol or the agreement is to ask you to contribute part of your computing power of your PC. To contribute part of computing power to the system when the blockchain technology is proposing blocks and achieving consensus. 

This process consumes a lot of computing power. So in decentralized data storage or decentralized computing system, it’s always been crucial that the protocol can decide which nodes will be included in current consensus proposing.

This involves a dynamic decision about which nodes to allocate this task to this time. So this is where AI can help.

AI keeps tracking all that data. Keep tracking, keep predicting, and keep summarizing data. So ideally, AI should be very capable of measuring the efficiency of each node.

For example, if I have a consensus task, then what are the nodes I should allocate to? I think this kind of decision is what AI is really good at. 

I think for all those decentralized processes on the consensus layer and the data management layer, data storage layer, AI can help in making the decision.”

[Q]: How does this decentralized nature of blockchain complement the analytical capability of AI?

[Vivien Lin]: “I’ll give you an example. One of the trending sectors in the blockchain side or technology side is called zero knowledge proof.

This ZK technology is quite a cutting edge approach to enhance trust and privacy in various industries. Actually, as far as I know, this technology has been implemented in national security. In many of those very confidential and important nationalized projects.

But it is also implemented in normal life cases such as investment or asset management verification. 

For example, if an asset manager claims to adhere to a specific investment strategy. If you invest in a quantitative fund or a hedge fund, their manager always tells you, ‘I invest in this sector and not that sector. I will allocate my assets no more than 5% in each stock.’

But actually, when their strategy becomes more complicated, it’s very hard for people to track or to verify if they really stick to the strategy that they claim to.

Using AI-enhanced ZK proof is a way that enables an investor to verify that their manager adheres to the strategy they claim without really revealing the strategy’s confidential details. 

For example, if they trained a quantitative model, right? Basically, they cannot reveal how they put the weights and how they automatically adjust the weights.

Especially if the strategy itself is also designed by certain AI model, there is no chance that they will reveal all the details. 

But how can we verify that they adhere to what they claim? Now we can use ZK knowledge and especially AI-enhanced ZK knowledge.

I think this kind of application or this kind of use case is somewhere where blockchain technology can be used in a wide spectrum of cases.”

[Q]: How is modular blockchain being used with AI to enhance the scalability and efficiency of the way these transactions happen?

[Vivien Lin]: “Modular blockchain technology was just as I described, you separate the computing power, separate the storage space, make them decentralized. 

This is the simplest way to understand modular technology. And AI algorithm can manage and optimize the sharding process in modular blockchains.

So sharding splits the blockchain into smaller or more manageable pieces. So each module can process in transaction independently. So AI can help dynamically adjust how transactions are allocated.

And AI can also predict the transaction volume and adjust the system dynamically. For example, if you anticipate now is a high load period an AI system can scale, can proactively scale resources or reallocate transactions across different modules to maintain the performance without the Mainnet really blocking. 

So this is something that AI plus modular technology can enhance the overall speed of transaction in blockchain transactions. And also AI can assist the optimization of the execution of smart contracts.

This is still in an experimental phase, I would say. As far as I know, not many smart contracts really use AI to predict the path or outcomes based on the historical data. But this is definitely something I know many people want to do.”

[Q]: You mentioned some of the solutions are very experimental at the moment. If you look towards the future, what potential developments do you see in the integration of AI and blockchain that is going to change the trading landscape?

[Vivien Lin]: “I think, one thing is trustless AI service. Blockchain can enhance trust in the AI decision-making process by making it transparent and verifiable.

And smart contracts could be used to validate AI decisions before any transaction or trade is executed. This is, I think, how smart contracts or how decentralized solutions meet AI technology. 

We’re always talking about how AI could aid blockchain, but here is how smart contracts or blockchain technology can aid AI decisions.

This is one thing. And another is cross-chain analysis. AI could manage and analyze data across multiple chain platforms.

Not many people have the required knowledge or required skill to access the data on the chain. Because different mainnet may have different coding language. So AI can help people to with just one click to get all that information from different mainnet.

And second is once you get all those data, how to analyze that. This is how we’re trying to use AI to detect fraud as well. If we rely on humans to do that, then that’s almost impossible, especially with current cyber attacks, the hack technology also improved. 

Now it’s almost impossible for human to do that. So I think AI could definitely aid on that space.”

Blockchain and AI are increasingly being used as complementary technologies to change how we transact, interact online, and do business.

You can learn more about how innovative tools and features are making trading crypto easier by contacting Vivien Lin or heading over to the BingX trading platform.

The post Interview: Vivien Lin, ​​Chief Product Officer at BingX appeared first on DailyAI.