Title of the talk:  EVT Based Rate-Preserving Distributional Robustness for Tail Risk
Abstract: Risk measures such as Conditional Value at Risk (CVaR) focus on extreme losses, where limited data makes model errors unavoidable. To hedge against model misspecification, practitioners evaluate worst-case tail risk under ambiguity sets.  Working with a fairly general class of distortion risk measures, we use Extreme Value Theory (EVT) to characterize the asymptotic scaling of worst-case tail risk under standard ambiguity sets, including Wasserstein balls and classical Φ-divergence neighborhoods, and show that robustification can inflate tail risk at a different first order EVT rate than the nominal model as the tail level 𝛃 → 0. Motivated by this diagnostic, we propose a tail-calibrated robustification - Rate-Preserving EVT-DRO (RPEV-DRO), designed to match the nominal model’s asymptotic rate for risk while still guarding against misspecification. Under standard domain of attraction assumptions, we prove that the worst-case risk under RPEV-DRO grows at the same first-order asymptotic rate as the baseline as 𝛃 → 0 , uniformly over key tuning parameters. We also establish analogous guarantees for a data-driven implementation based on consistent tail-index estimation. Synthetic and real data experiments, including network style systemic risk losses, show that RPEV-DRO avoids the severe risk inflation often induced by Wasserstein and Φ-divergence formulations. 

Title of the talk:  Random Neural Network Algorithm for Solving Nonlinear PDEs in High-Dimensional Option Pricing (Online)
Abstract: We present a random neural network–based algorithm that efficiently solves high-dimensional nonlinear partial differential equations (PDEs) and apply it to the pricing of high-dimensional financial derivatives under default risk. Our empirical results demonstrate that the algorithm can approximately solve nonlinear PDEs in up to 10,000 dimensions within seconds.

This talk is based on joint works with Christian Beck, Sebastian Becker, Patrick Cheridito, and Arnulf Jentzen, as well as Philipp Schmocker and Sizhou Wu.

Title of the talk:  E-commerce sellers' ratings and FinTech innovations
Abstract:  Coming soon!

Title of the talk: Rollover Risk and Systematic Risk of Shadow Banks
Abstract: We investigate the relation between rollover risk and systemic risk exposure in the shadow banking sector. Our sample consists of hand collected data from annual reports of retail non-bank financial companies (Retail-NBFCs), which constitute a significant part of the shadow banking sector in India, and commercial paper holdings of liquid debt mutual funds (LDMFs), which are a key source of financing for Retail-NBFCs. In 2018, Retail NBFCs experienced a sharp increase in rollover risk following defaults by subsidiaries of Infrastructure Financing and Leasing Services (IL&FS). We use this empirical setting to construct a rollover risk measure and show that it is more informative about systemic risk exposure than existing proxies. We also show that systemic risk exposure of Retail-NBFCs is positively related to our measure of rollover risk and negatively related to the level of liquidity buffer in the LDMF sector; furthermore, we show that the impact of rollover risk on systemic risk exposure is less adverse when liquidity buffer in the LDMF sector is high.

Title of the talk:  Sustainable Investment: Maintaining portfolio performance while reducing the carbon footprint
Abstract: Amid the global crisis of climate change, urgent action is imperative. In this study, we develop two types of decarbonized indices, which render a dynamic hedging approach for passive investors. Focusing on long-term returns with minimal active trading and risk exposure, we create the decarbonized indices for NIFTY-50, a benchmark index for the Indian market. Proposed methodology relies on suitable optimization techniques to choose the portfolio weights that minimize the tracking error while significantly reducing carbon footprints. These indices are shown to perform better than existing benchmarks, especially during major climate events. They are likely to offer investors a buffer to adapt to climate policies and carbon pricing. In ongoing work, we use tail risk measures instead of variance to control the risk of the portfolio.

Title of the talk: A Primer on Blended Finance and its Framework
Abstract:  In this expository lecture, we present a structured overview of Structured Blended Finance (SBF) design, with an emphasis on the quantitative drivers of risk management and cash flow distribution. We begin with the general architecture of blended finance, highlighting the role of concessional capital within the paradigm of asymmetric payoff profiles for heterogeneous investor objectives. Traditional credit risk modelling approaches, including intensity models, default probabilities and dependency structures, are used in analyzing the portfolio loss distributions. These loss distributions are pertinent in determining the characteristics of the tranches. We then move on to examining the cash flow framework of pay-through structures, highlighting how contractual waterfalls translate portfolio-level outcomes into tranche-specific returns over the life of the fund. Finally, an algorithm on the operational aspects of SBFs is discussed to demonstrate how credit risk modelling and simulated cash flows are combined to evaluate tranche risk-return profiles and to calibrate fund structures, in a manner that is consistent with senior investor constraints in conjunction with the objectives of the originator.

Title of the talk: A Systems Lens on Emerging Financial Frontier
Abstract: Traditional model-centric thinking may no longer be enough, when financial systems are becoming increasingly complex and interconnected. This talk explores how applying a systems thinking lens can unlock new value across emerging areas in finance. We will try and understand the challenges such as:

What happens when supply chains are reimagined through digital tokenization?
When autonomous AI agents begin transacting on behalf of humans—and how do we model such ecosystems?

Title of the talk: What needs to change in DeFi for wider adoption?
Abstract: Decentralized Finance (DeFi) has introduced a new financial paradigm by enabling permissionless, programmable financial services without centralized intermediaries. While DeFi protocols demonstrate strong innovation and composability, their  broad based adoption requires some issues to be resolved.  This talks  examines key barriers to broader DeFi adoption, focusing on privacy and usability challenges that constrain both retail and institutional participation. Public transaction transparency exposes sensitive financial information, creating confidentiality and compliance concerns, while complex user experiences—spanning wallet management, transaction finality, and fragmented interfaces—raise the barrier to entry for non-technical users.

We argue that overcoming these challenges requires a shift from protocol-centric design toward user- and institution-aware architectures. Privacy-preserving techniques such as zero-knowledge proofs, selective disclosure, and confidential execution can enable secure yet compliant interactions. In parallel, improved abstraction layers and execution environments are needed to simplify user interaction with smart contracts. We conclude by outlining research directions that align cryptographic innovation with usability and real-world financial integration.

Title of the talk: Recent trends of practical areas of research in investments and risk
Abstract: In an era dominated by the ubiquity of Smart Beta, the traditional boundary between skill-based returns and systematic factors has blurred. This plenary session explores the critical evolution of modern portfolio construction, beginning with a provocative interrogation of ""Pure Alpha"": Is it a genuine extraction of unique value, or merely a disguised beta captured by increasingly sophisticated factor models?

By decomposing returns through advanced econometric ""drill-down"" techniques, we isolate the residual drivers of performance, stripping away the noise of alternative risk premia to find what truly remains of idiosyncratic skill. However, the path to alpha is fraught with cognitive and structural traps. We will revisit the ""Seven Sins of Quantitative Investing""—ranging from backtest overfitting to the neglect of transaction costs—to understand why even the most robust models often fail upon contact with live markets.

Finally, the discussion transitions from individual security selection to risk management and sneak into latest industry trends on Total Portfolio Approach (TPA), a holistic framework designed to build portfolios.

Title of the talk: ML for Pattern Discovery in Equities
Abstract: This talk provides a practical overview of how machine learning techniques can be applied to discover patterns and predictive signals in equity markets. After introducing core ML paradigms—supervised, unsupervised, and reinforcement learning—we walk through the end-to-end data-science pipeline used in quantitative trading: data preparation, feature engineering, model selection, and robust evaluation. Special emphasis is placed on challenges unique to financial data, including noise, non-stationarity, and overfitting, along with methods such as regularization, cross-validation, and ensemble approaches to address them. The session concludes with real-world use cases—returns prediction, regime modeling, peer discovery, anomaly detection—and research directions involving deep learning, graph networks, and reinforcement learning, for strategy design. 

Title of the talk: A new approach for the pricing of share buyback contracts
Abstract: In this talk, we will review a recent methodology for pricing share buyback contracts that replaces traditional control-based approaches with optimized heuristic strategies designed to maximize contract value. The valuation framework builds on classical techniques used for pricing path-dependent Bermudan options, enabling efficient numerical implementation. An additional feature of the approach is that it naturally leads to a corresponding hedging strategy. The material presented in this talk is based on recent work by Bastien Baldacci and co-authors.

Title of the talk: New Rails, New Instruments, New Participants: What Happens When They Share an Execution Layer
Abstract: Stablecoins have matured into programmable settlement rails that are fast, global, and increasingly used at scale. Tokenization is producing instruments that carry their own compliance, transfer, and settlement logic natively, making them composable in ways traditional instruments aren't. And a new category of participant is arriving: autonomous AI agents that can hold assets and transact independently. Right now, they have no identity, no reputation, and no way to pay each other without going through existing accounts and rails that weren't built for them. ERC-8004 gives agents portable identity and reputation on-chain. x402 lets them settle payments in a single HTTP request at near-zero cost.

Separately, these are well-understood developments. Together, on shared infrastructure, they start to compound. An agent can settle against a tokenized instrument, accumulate reputation from the interaction, and use that reputation to access better terms on the next one, all within a single composable environment, with no intermediary in the loop. The dynamics that emerge, new forms of counterparty assessment, continuous settlement, machine-scale liquidity demands, programmable compliance that travels with the asset rather than being enforced externally, don't arise when these layers operate apart. How these protocols connect at the execution layer, what assumptions they encode, and what they make possible when they interact is where the most interesting and least examined design space sits today.

Title of the talk: What needs to change in DeFi for wider adoption?
Abstract: Decentralized Finance (DeFi) has introduced a new financial paradigm by enabling permissionless, programmable financial services without centralized intermediaries. While DeFi protocols demonstrate strong innovation and composability, their  broad based adoption requires some issues to be resolved.  This talks  examines key barriers to broader DeFi adoption, focusing on privacy and usability challenges that constrain both retail and institutional participation. Public transaction transparency exposes sensitive financial information, creating confidentiality and compliance concerns, while complex user experiences—spanning wallet management, transaction finality, and fragmented interfaces—raise the barrier to entry for non-technical users.

We argue that overcoming these challenges requires a shift from protocol-centric design toward user- and institution-aware architectures. Privacy-preserving techniques such as zero-knowledge proofs, selective disclosure, and confidential execution can enable secure yet compliant interactions. In parallel, improved abstraction layers and execution environments are needed to simplify user interaction with smart contracts. We conclude by outlining research directions that align cryptographic innovation with usability and real-world financial integration.

Title of the talk: Static Factors to Adaptive Signals: The Evolution of Quantitative Investing
Abstract: Quantitative investing has evolved rapidly with advances in data availability, modelling techniques, and computational scale. This talk discusses how classical investment factors—such as value, quality, and sentiment—are being augmented through machine-learning-based signal extraction, alternative data sources, and event-driven indicators. Drawing on practical examples from global equity and fixed-income markets, the session highlights the limitations of static factor and smart-beta approaches, particularly in emerging markets, and motivates the shift toward dynamic factor construction. The talk also explores the growing role of systematic fixed-income strategies enabled by improved market transparency and electronic trading, and reflects on the research infrastructure required to support scalable, adaptive quantitative investment frameworks.

Title of the talk: Market Microstructure in Decentralized Finance : Design, Latency, and Strategies
Abstract: Decentralized Finance (DeFi) has progressed from simple automated market makers to on-chain markets with increasingly sophisticated microstructure. This talk examines how foundational ideas from quantitative finance—market design, price discovery, latency, and risk management—are being reimplemented in blockchain-based systems, and the implications for systematic and high-frequency trading. We analyze the dominant DeFi market architectures, contrasting pool-based designs built on automated market makers and concentrated liquidity with central limit order book (CLOB)–based exchanges. The discussion highlights how pricing functions, liquidity placement, and protocol constraints shape the feasible strategy space for quantitative traders. Particular attention is given to recent CLOB-based systems that demonstrate low-latency, on-chain execution through custom blockchain design and Byzantine fault-tolerant consensus. The talk further situates decentralized markets relative to centralized exchanges, emphasizing the trade-offs between transparency, latency, and custodial trust. Finally, we explore the emerging role of network- and hardware-level optimization, arguing that the next phase of quantitative finance in DeFi will be driven not only by financial modeling, but also by advances in distributed systems, networking, and hardware-aware market design.

Title of the talk: Toward Intelligent Financial Decision-Making: Decision-Centric AI, Interactive Optimization, and Quantum Computing
Abstract: Quantitative finance is undergoing a structural shift from prediction-driven workflows toward intelligent systems that directly operationalize financial decision-making. This talk presents three research thrusts advancing this transition.

We first examine decision-focused learning with deep neural architectures that generate optimized decision variables, challenging the traditional predict-then-optimize paradigm. Next, we introduce cognitive agent architectures that leverage multi-agent large language model systems for financial analysis, personalized portfolio construction, and interactive refinement. Finally, we explore quantum and quantum-inspired optimization for asset allocation and portfolio optimization, investigating their quantum potential for improved sampling efficiency and scalable approximation. Collectively, these directions signal a move beyond passive modeling toward adaptive, decision-centric intelligence in quantitative finance.

Title of the talk: Counterparty Credit Risk of Traded Derivatives
Abstract: Counterparty Credit Risk (CCR) refers to the risk of financial loss arising from a counterparty’s potential default before the maturity of a derivative contract. Unlike traditional credit risk, CCR is bilateral and driven by the stochastic evolution of market variables such as interest rates, foreign exchange rates, and equity prices, leading to time-varying and path-dependent exposures. Effective CCR measurement relies on exposure metrics including Expected Exposure (EE), Potential Future Exposure (PFE), and valuation adjustments such as Credit Valuation Adjustment (CVA). Modern regulatory frameworks emphasize accurate CCR modelling, particularly incorporating dependencies between exposure and credit quality, including wrong-way risk. Collateralization, netting, and central clearing play key roles in mitigating CCR, while advanced numerical methods such as Monte Carlo simulation are widely used for valuation and risk management.

More recently, Artificial Intelligence and Machine Learning techniques are gaining significant attention in CCR modelling, offering potential improvements in computational efficiency, exposure approximation, scenario generation, and real-time risk analytics, thereby complementing traditional quantitative approaches.

Title of the talk: Toward Intelligent Financial Decision-Making: Decision-Centric AI, Interactive Optimization, and Quantum Computing
Abstract: Financial systems are moving past prediction-driven workflows towards intelligent systems that operationalize decision-making. This talk presents three research thrusts advancing this transition. 

We first examine decision-focused learning with deep neural architectures that generate optimized decision variables, challenging the traditional predict-then-optimize paradigm. Next, we introduce cognitive agent architectures that leverage multi-agent large language model systems for financial analysis, personalized portfolio construction, and interactive refinement. Finally, we explore quantum and quantum-inspired optimization for asset allocation and portfolio optimization, investigating their quantum potential for improved sampling efficiency and scalable approximation. 

Collectively, these directions signal a move beyond passive modeling toward adaptive, decision-centric intelligence in quantitative finance.