2018 spotlight: ML/AI
GS was among the pioneers to replace human traders with AI. At its height back in 2000, the US cash equities trading desk at Goldman Sachs’s New York headquarters employed 600 traders, buying and selling stock on the orders of the investment bank’s large clients. Today there are just two equity traders left. Automated trading programs have taken over the rest of the work, supported by 200 computer engineers. The pattern will be repeated over and over, with 'Warren in a Box' being the most recent example.
'Warren in a Box' to Pick Stocks for First Nordic AI Fund
- Finnish fund manager FIM is introducing the first investment fund in the Nordic region, where a self-learning algorithm gets to pick all the stocks.
- The FIM Artificial Intelligence fund seeks to tease out patterns even an experienced fund manager may not detect. “This is the next revolution across society, including in investing,” Chief Investment Officer Eelis Hein, who oversees $6.6 billion in investments at FIM Asset Management, said.
- The “Warren in a box” technology, was in development for >2 years.
- The algorithm picks 50 stocks from a pre-screened universe of 4,000 liquid, developed-market equities, each with a market capitalization of 1 billion euros or more. After running the Warren algorithm, the AI applies a team-building algorithm, churning out a suitably diversified portfolio from the individual stock picks.
- The database contains fundamental data on 700 variables going back to 1986. The fund’s allocation takes place every six months and the actual trading is done by a human.
How Verizon is Building a Big Data and AI Culture
- Verizon Communications is a massive telecommunications conglomerate with $126 billion in 2016 revenues.
- Verizon has a variety of different analytics and AI groups scattered around the company.
- Verizon's Data Science and Cognitive Intelligence (DSCI) group focuses on applying analytics and cognitive technology to Verizon’s interactions with customers. DSCI is heavily focused on cognitive technologies and is adding increasing levels of intelligence to the company’s marketing and customer service applications.
- Global Supply Chain Strategy and Analytics group ensures that products reach customers and that sourcing and procurement are effective. The group is increasingly knowledgeable and proficient in ML applications.
- Big Data and Artificial Intelligence Systems, focuses on creating new products and services for Verizon with these methods and tools. The group employs large-scale ML both to improve Verizon’s infrastructure and to develop products in such areas as education, healthcare, and IoT.
- Verizon Enterprise Data Analytics (VEDA) is an enterprise organization that addresses data management, data governance, data warehousing and data lakes, and common analytical and AI technologies. The goal of VEDA is to facilitate cross-functional, cross-organizational projects.
- Verizon has both proprietary and open source analytical and AI technologies.
- Having a variety of different groups with a coordinating mechanism like VEDA appears to offer an effective organizational solution to the challenge and opportunity of leveraging big data and AI at Verizon.
Why Decentralized Artificial Intelligence Will Reinvent the Industry As We Know It
- We are moving toward the next frontier – decentralized AI that can run and train on local devices or make decisions in decentralized networks like blockchain.
- The transition to decentralized AI is enabled by new technologies, such as Google’s Federated Learning, that allow for crowd-training of ML algorithms, device-centric AI that runs and trains ML models on mobile devices and the use of AI in DAOs (decentralized autonomous organizations) on blockchain networks.
- Powered by data flowing from thousands of users, and having access to resources and the ability to amass them, decentralized AIs can become a source of huge economic value for its owners.
- Using generative models (GANs), we can create AI DAOs that trade in their own art, logos, sketches, images or video clips and distribute profits as cryptocurrency tokens to their shareholders.
- AI DAO can become the only shareholder of the accumulated capital. Example: Terra0. , a project involving an augmented self-owned forest proposed by Paul Seidler and Paul Kolling from the University of Arts, Berlin.
- Centralized AI solutions provided as APIs and cloud-based services have certain bottlenecks. Since users access AI features via the network and because ML algorithms involve heavy computations, high latency is often an issue. Also, if you train AI models in a centralized way, it may take more time to improve them.
- Decentralized AI can function locally on users’ devices, have access to more user data and have no dependence on a network connection, which means less power consumption and minimal latency. Recent advances in decentralized AI have been made thanks to on-device optimization of AI/ML for smartphones and production of dedicated chips for mobile AI and for desktops (e.g., Google’s TPU).
The Monetary Authority of Singapore (MAS) announced details of the new $27million Artificial Intelligence and Data Analytics (AIDA) Grant under the Financial Sector Technology and Innovation (FSTI) Scheme
- The new grant aims to promote the adoption and integration of AI and data analytics in financial institutions.
- The AIDA Grant will focus on two tracks: the Financial Institution Track and the Research Track.
- Under the Financial Institution Track, the AIDA Grant will co-fund up to 50% of project costs for Singapore-based financial institutions which leverage AI and data analytics techniques to generate insights, formulate strategy, and assist in their decision making. These techniques may include ML, NLP or text analytics, deep learning or neural networks, predictive and prescriptive analytics.
- Under the Research Track, the AIDA Grant will co-fund research institutions’ AI or data analytics projects which have clear applications for Singapore’s financial sector. The AIDA Grant will provide up to 70% co-funding for eligible projects.
- The Grant will also provide funding for research projects submitted through periodic calls for proposals on specific AI or data analytics topics that benefit the industry. Such calls will be published on the MAS website.
- Applications are now open for both the Financial Institution Track and Research Track.
RBC readies to launch robo-advisor platform
- RBC test ran its RBC InvestEase platform.
- The new business will offer automated investment advice and discretionary portfolio management to consumers through a digital interface supported by accredited portfolio advisors.
Paytm and ICICI Bank tie-up to offer short-term instant digital credit
- Paytm has partnered with ICICI Bank to launch ‘Paytm-ICICI Bank Postpaid’ to offer access to interest-free short-term digital credit.
- This is the country’s first tie-up between a Scheduled Commercial Bank and a payments platform to offer instant digital credit for everyday use-cases.
- Paytm-ICICI Bank Postpaid is a digital credit account with instant activation: with no hassles of documentation or branch visit. Activation is fully online. There is no transaction joining or hidden administration fees either.
- Available 24x7, it is based on a new Big Data-based algorithm by ICICI Bank for the real-time credit assessment of customers. The algorithm uses an intelligent combination of the financial and digital behavior of the customer including credit bureau check, purchase patterns, frequency of purchase to ascertain the creditworthiness of a customer within a few seconds.
- Based on the credit score of the customer, the bank offers up to 45 days interest-free credit limit. It ranges from Rs. 3,000 to Rs. 10,000, extendable up to Rs. 20,000 based on the repayment history. Paytm-ICICI Bank Postpaid will also offer a quick checkout to customers with the Paytm Passcode.
- Customers can use their Paytm Wallet, debit card or internet banking of any bank for an easy repayment of their dues.
PayPal to sell $6 billion in consumer loans to Synchrony Financial
- PayPal agreed to sell $5.8 billion in consumer credit receivables to Synchrony Financial. These loans will generate $1 billion in revenue for 2017.
- The deal also includes Synchrony’s acquisition of $1 billion in participation interests in PayPal receivables held by certain investors and a chartered financial institution.
- As a result of the deal, the two companies will expand their partnership by making Synchrony Bank the exclusive issuer of the PayPal Credit online consumer financing program available to PayPal customers in the US for the next 10 years.
- For PayPal, the deal means it will lose the interest the loans generate, but it will free up billions that it can use to grow its business in other ways.
- Synchrony benefits by tapping into a source of consumer loans that’s more tied to online shopping – which is increasingly where consumers make large purchases requiring credit.
American Express, Santander team up with Ripple for cross-border payments via blockchain
- American Express and Santander have partnered with Ripple to speed up cross-border payments between the US and the US by using blockchain technology.
- Payments made by American Express' business customers on its FX International Payments (FXIP) platform will now be routed through Ripple's enterprise blockchain network, RippleNet.
- "This collaboration with Ripple and Santander represents the next step forward on our blockchain journey, evolving the way we move money around the world," said Marc Gordon, EVP and chief information officer at American Express.
- American Express' blockchain project will initially allow customers in the US to connect instant, traceable cross-border non-card payments to UK Santander bank accounts.
Fujitsu Launches Inter-Blockchain Payment System
- Fujitsu launched ConnectionChain payments technology aimed at facilitating transactions between blockchain in mid-November 2017. The payments system intends to provide a means for different cryptocurrency networks to interoperate with each other.
- To effectively achieve a successful settlement between digital currencies, there should be a dependable application to manage the currency exchange processing at the boundaries between the blockchains in a transparent manner.
- The payments technology can resolve the issues surrounding such transactions by inter-connecting multiple blockchains and execute a single transaction process that can be automatically executed.
- "Fujitsu Laboratories has now developed an extension of smart contract technology which inter-connects multiple Blockchains by recording the series of related transactions on each chain in a dedicated Blockchain, or a "connection-type chain," to link to the currency exchange into a single transaction process that can be automatically executed. It has also developed a transaction control technology to synchronize execution timing of the transaction process on each chain."
- Fujitsu plans to commercialize the technology starting 2018 (depending on the results of tests).
On InsurTech (reality check)
- About 95% of investment in InsurTech distribution models will be wasted, claims Guidewire Chief Executive Marcus Ryu. Recent investments in InsurTech amount to a bubble. $3.5bn has been invested in InsurTech over the past few years. It is not possible to spend that money productively, Ryu warned.
- Areas of InsurTech spending that are relevant and useful: satellite and commercial sensor data aggregation and photo-based estimating.
- Attempts to change insurance distribution models will not be a disruptive force and worth the amount of money that has been funneled into it.
- InsurTech companies are using the metric of customer acquisition to show progress. It is a good way to measure progress for consumer internet firms but does not lead to a tangible benefit for an insurer. “If you’re an insurance company it’s very easy to get customer acquisition. Any insurance company can acquire all the customers they want and lose money on every one of them. The fact that they were thinking in terms of consumer internet attitude [is telling], you have to spend a lot, take lot of losses, but acquire a lot of customers and then figure out how to make money on them later. It does not work.”
- Interesting things were being seen in smaller volume markets. Success there is likely caused by the ability to tackle a small segment of the overall market with niche products.
- P2P insurance risk-sharing models are tiny and not consequential yet. They should be viewed as experiments and they are all currently “significantly loss-making.”
*Featured image credit: Chatbots Magazine.