A successful programme of personalised discounts and recommendations relies on identifying products that customers want, based both on items bought in the past and on relevant products that the customers have not yet purchased. Using basket-level grocery shopping data, we aim to use clustering ("community detection") techniques to identify groups of shoppers with similar preferences, along with the corresponding products that they purchase, in order to design better recommendation systems.

Stochastic block models (SBMs) are an increasingly popular class of methods for community detection. In this talk, I will expand on some work done by Newman and Clauset [1] that uses a modified SBM for community detection in annotated networks. In these networks, additional information in the form of node metadata is used to improve the quality of the inferred community structure. The method can be extended to bipartite networks, which contain two types of nodes and edges only between nodes of different types. I will show some results obtained from applying this method to a bipartite network of customers and products. Finally, I will discuss some desirable extensions to this method such as incorporating edge weights and assessing the relationship between metadata and network structure in a statistically robust way.

[1] Structure and inference in annotated networks, MEJ Newman and A Clauset, Nature Communications 7, 11863 (2016).

Note: This talk will cover similar topics to my presentation in the InFoMM group meeting on Friday, November 25 but it won't be exactly the same. I will focus more on the mathematical details for my JAMS talk.
 

This talk will introduce the notion of quasi-convex subgroups. As an application, we will prove that the intersection of two finitely generated subgroups of a free group is again finitely generated.
 

Cover’s celebrated theorem states that the long run yield of a properly chosen “universal” portfolio is as good as the long run yield of the best retrospectively chosen constant rebalanced portfolio. We formulate an abstract principle behind such a universality phenomenon valid for general optimization problems in the long run. This allows to obtain new results on modelfree portfolio optimization, in particular in continuous time, involving larger classes of investment strategies. These modelfree results are complemented by a comparison with the log-optimal numeraire portfolio when fixing a stochastic model for the asset prices. The talk is based on joint work with Walter Schachermayer and Leonard Wong.

Thomas Piketty's influential book “Capital in the Twenty-First Century” documents the marked and unequivocal rise of income and wealth inequality observed across the developed world 
in the last three decades. His extrapolations into the distant future are much more controversial and has 
has been subject to various criticisms from both mainstreams and heterodox economists. This motivates the search for an alternative standpoint incorporating 
heterodox insights such as endogenous money and the lessons from the Cambridge capital controversies. We argue that the Goodwin-Keen approach paves the road towards such an alternative.
We first consider a modified Goodwin-Keen model driven by consumption by households, instead of investment by firms, leading to the same qualitative features 
of the original Keen 1995 model, namely the existence of an undesirable equilibrium characterized by infinite private debt ratio and zero employment, 
in addition to a desirable one with finite debt and non-zero employment. By further subdividing the household sector into workers and investors, we are able to investigate their relative 
income and wealth ratios for in the context of these two long-run equilibria, providing a testable link between asymptotic inequality and private debt accumulation.

We consider impatient decision makers when their assets' prices are in undesirable low regions for a significant amount of time, and they are risk averse to negative price jumps. We wish to study the unusual reactions of investors under such adverse market conditions. In mathematical terms, we study the optimal exercising of an American call option in a random time-horizon under spectrally negative Lévy models. The random time-horizon is modeled by an alarm of the so-called Omega default clock in insurance, which goes off when the cumulative amount of time spent by the asset price in an undesirable low region exceeds an independent exponential random time. We show that the optimal exercise strategies vary both quantitatively and qualitatively with the levels of impatience and nervousness of the investors, and we give a complete characterization of all optimal exercising thresholds. 

The share of market making conducted by high-frequency trading (HFT) firms has been rising steadily. A distinguishing feature of HFTs is that they trade intraday, ending the day flat. To shed light on the economics of HFTs, and in a departure from existing market making theories, we model an HFT that has access to unlimited leverage intraday but must fund any end-of-day inventory at an exogenously determined cost. Even though the inventory costs only occur at the end of the day, they impact intraday price and liquidity dynamics. This gives rise to an intraday endogenous price impact mechanism. As time approaches the end of the trading day, the sensitivity of prices to inventory levels intensifies, making price impact stronger and widening bid-ask spreads. Moreover, imbalances of buy and sell orders may catalyze hikes and drops of prices, even under fixed supply and demand functions. Empirically, we show that these predictions are borne out in the U.S. Treasury market, where bid-ask spreads and price impact tend to rise towards the end of the day. Furthermore, price movements are negatively correlated with changes in inventory levels as measured by the cumulative net trading volume.
 

(based on joint work with Tobias Adrian, Erik Vogt, and Hongzhong Zhang)

We propose a new flexible unified framework for studying the time consistency property suited for a large class of maps defined on the set of all cash flows and that are postulated to satisfy only two properties -- monotonicity and locality. This framework integrates the existing forms of time consistency for dynamic risk measures and dynamic performance measures (also known as acceptability indices). The time consistency is defined in terms of an update rule, a novel notion that would be discussed into details and illustrated through various examples. Finally, we will present some connections between existing popular forms of time consistency. 
This is a joint work with Tomasz R. Bielecki and Marcin Pitera.