Driverless truck startup Starsky Robotics is shutting down, but not before sharing some cold hard truths about the autonomous driving industry. Founded in 2015, Starsky proposed a combination of self-driving and remote control for a fleet of next-generation trucks, saving full autonomy for the highway.

Rather than build a driverless truck that could handle every situation, Starsky’s plan was to mix autonomous systems with human teleoperation. On the highway, a relatively controlled environment, the truck would drive itself. That way, the demands for skilled operators would be significantly reduced.

In trickier situations, however – effectively “the first and last mile,” as Starsky explained it – a human driver would take over the controls. They wouldn’t be physically present in the truck, mind. Instead they’d use remote controls to pilot the rig from a distance.

Back in 2019, Starsky demonstrated the first fully-unmanned truck to drive on a live, public highway. Now, though, the company is shutting down. In a blunt post-mortem of what went wrong, founder and CEO Stefan Seltz-Axmacher blamed results-hungry investors, unexpected difficulties with getting the AI right, and the fact that safety just isn’t sexy for Starsky’s problems – and the problems that he predicts will impact the self-driving industry as a whole.

“There are too many problems with the AV industry to detail here,” Seltz-Axmacher writes, “the professorial pace at which most teams work, the lack of tangible deployment milestones, the open secret that there isn’t a robotaxi business model, etc. The biggest, however, is that supervised machine learning doesn’t live up to the hype. It isn’t actual artificial intelligence akin to C-3PO, it’s a sophisticated pattern-matching tool.”

The issue, he explains, is that matching – and eventually exceeding – human drivers’ abilities with edge cases is much tougher than most realized. Everyday driving in reasonable conditions is fairly low-hanging fruit; that can be achieved relatively rapidly. Developing a system that is capable of reacting safely to unexpected situations, however, is far trickier, and as you refine the self-driving AI you also set yourself the challenge of finding increasingly specific risk models with which to test.

Adding to that problem is the fact that, while safety is often cited as a primary concern for people when asked about whether they’d get into an autonomous vehicle, it’s actually a tough thing to get people excited about. The same, Seltz-Axmacher says, goes for investors. Starsky spent almost two years working on safety engineering, but “the problem is that all of that work is invisible,” he writes.

“Investors expect founders to lie to them,” the Starsky founder explains, “so how are they to believe that the unmanned run we did actually only had a 1 in a million chance of fatality accident? If they don’t know how hard it is to do unmanned, how do they know someone else can’t do it next week?”

At the end of 2019, the company’s attempts to raise more money fell flat. It’s currently seeking to sell off its patents as the company breaks apart. Seltz-Axmacher says he sees real autonomy still being 10 years out; “no one should be betting a business on safe AI decision makers,” he concludes.

The post Driverless truck startup Starsky Robotics folds: CEO shares tough autonomy home truths appeared first on Artificial Intelligence.

Gareth Coltman, global head of automation at MarketAxess, the electronic platform for fixed income trading and reporting, said the industry is going through a second revolution which will lead to radical changes in market structure.

Coltman told Markets Media: “The first revolution in electronic trading involved automating request for quotes and Open Trading. We are seeing the second revolution in electronic trading which has been enabled by the pre-trade data provided by CP+ and will lead to radical changes in market structure.”

Open Trading, MarketAxess’ all-to-all trading protocol, allows multiple parties in a network to come together to trade, rather than the traditional model of only banks supplying liquidity to the buy side.

Composite+ (CP+)

CP+ is the firm’s algorithmic pricing engine that uses artificial intelligence to price corporate bonds using a variety of data sources including public reports and proprietary MarketAxess data.

David Krein, global head of research at MarketAxess, told Markets Media that the firm experimented with machine learning for 18 months before the launch of CP+ in May 2017, and found it clearly made better sense of the data.

“After searching for tools to speed up and improve our process, we decided to use H20.ai to implement the necessary algorithms,” he added.

H2O.ai said in a statement last month that its platform provides open source artificial intelligence and machine learning capabilities to MarketAxess’ CP+.

The pricing engine’s algorithm consumes more than 200 features and produces an unbiased, two-sided market for 95% of the tradable universe which is updated every 15 to 60 seconds, depending on the liquidity of the instrument.

“The predicted prices of CP+ track traded levels very closely, and we aim for zero average difference between the two,” said Krein. “A real-time accurate pre-trade reference price for corporate bonds has not been available before.”

Sri Ambati, chief executive and founder at H2O.ai, told Markets Media that the firm’s open source platform can perform one billion regressions in less than five seconds.

“This ensures data is correct in rapidly changing markets, which is very powerful when combined with MarketAxess’s domain knowledge in fixed income,” said Ambati.

The technology provider also has a H2O Driverless AI platform which uses automation to accomplish tasks including model validation, model selection and deployment and machine learning interpretability much more quickly.

Ambani explained that testing algorithms automatically allows firms to “fail faster.” He added: “By testing with more rapid iterations the right strategy can be found more quickly.”

Krein continued that artificial intelligence has been used in CP+ for euro bonds and emerging markets globally.

“We have already extended the facility into eight local emerging markets, where data is hardest to come by, making a pricing tool such as CP+ that much more valuable,” he added.

Electronic trading 

In the US, the outperformance of electronic trading can be evidenced by comparing venue transactions to Trace, the reporting system.

Krein said: “In other markets, we have used CP+ as a benchmark as well and we can now observe, measure, and track the same results.”

MarketAxess has constructed trade performance indexes against Trace as a benchmark. The MarketAxess U.S. Investment Grade Trading Performance Indexes give a view of the relative “over-performance” or “under-performance” associated with venue selection.

When the indexes were launched in April last year MarketAxess said in a statement that participants had an estimated improvement of 0.9 basis points in yield per bond in March 2019 when trading US Investment Grade credit on the MarketAxess platform versus trading away.

“We will be rolling it out to other markets in the coming months,” Krein added.

Automation

Coltman said: “Fixed income currently has automation of traditional RFQs but new tools and protocols are emerging which will impact all trading activity.”

For example, MarketAxess launched Live Markets for new issues and portfolio trading six months ago. Live Markets is a protocol for Open Trading which creates a single view of two-way, actionable prices for the most active bonds. In addition, executing in Live Markets gives evidence of best execution as the platform provides details of the quotes in the market at that point in time.

Coltman continued that transactions are ripe for automation and adoption will soon become ubiquitous in fixed income.

“We only launched automated execution two years ago and already between 50% and 60% of activity for some of our biggest clients is auto-executed,” he added. “The industry does not realise how fast this is happening.”

AI in financial services

Ambani said: “We are in the earliest days of the AI era in financial services.”

In December last year Credit Suisse selected H2O.ai as a member of its 2019 Disruptive Technology Recognition Program. The scheme is a joint initiative between the bank’s investment banking and capital markets division and the group chief technology officer function.

Ambati said: “H2O.ai was honored to be selected into the coveted Credit Suisse’s DTR program to partner across every group within the bank and co-invent AI.”

The partnership with the bank began when Credit Suisse adopted H2O Open Source in core finance and banking. Subsequently in 2019 the bank chose H2O Driverless AI to accelerate AI adoption in front-office and back-office in global markets, fixed income and capital markets.

In August last year another bank, Goldman Sachs, led a $72.5m series D ending in H2O.ai alongside the Ping An Global Voyager Fund,  which took total funding to $147m.  Jade Mandel from Goldman Sachs joined the H2O.ai board.

Erdit Hoxha, head of European equity trading at Goldman Sachs Securities Division, said in a statement at the time: “The results we’ve got with H2O are promising, we are now looking at wider adoption of the AI models across the equity trading floor for market making.”

Ambati continued that H2O.ai was founded in 2012 with the aim of democratizing AI for everyone.

“We want to provide signal sharing as a service and intelligence as a service,” he added. “The democratization of technology and talent will transform the whole capital markets industry and lead to more transparency and easier access to capital.”

The post ’Second Revolution’ In Electronic Bond Trading appeared first on Artificial Intelligence.