arXiv subjects

Mahsa Bastankhah

Publications and source records attributed to Mahsa Bastankhah.

7 recordsLinked to original sources

arXiv2026

Learning to Perceive the World Through Control: Empowerment-Based Representation Learning

In many practical reinforcement learning environments, observations are far higher-dimensional than the variables that matter for control. In this work, we ask: can we learn representations that capture only control-relevant features of the environment? We study this question through the empowerment objective, which maximizes an agent's influence over the environment and is widely used for unsupervised skill learning. We show that empowerment agents induce two distinct representations -- forward and backward -- that capture complementary aspects of the state, and both of which are invariant to control-irrelevant features. Thus, empowerment maximization leads agents to learn an implicit, control-centric model of the world. Our analysis highlights the importance of learning representations through interaction rather than from passive datasets: interaction aimed at maximizing control is essential for learning useful invariance properties, a perspective that aligns closely with the causal learning literature.

cs.LG↗

arXiv2025

Demystifying the Mechanisms Behind Emergent Exploration in Goal-conditioned RL

In this work, we take a first step toward elucidating the mechanisms behind emergent exploration in unsupervised reinforcement learning. We study Single-Goal Contrastive Reinforcement Learning (SGCRL), a self-supervised algorithm capable of solving challenging long-horizon goal-reaching tasks without external rewards or curricula. We combine theoretical analysis of the algorithm's objective function with controlled experiments to understand what drives its exploration. We show that SGCRL maximizes implicit rewards shaped by its learned representations. These representations automatically modify the reward landscape to promote exploration before reaching the goal and exploitation thereafter. Our experiments also demonstrate that these exploration dynamics arise from learning low-rank representations of the state space rather than from neural network function approximation. Our improved understanding enables us to adapt SGCRL to perform safety-aware exploration.

cs.LG↗

arXiv2025

AgileRate: Bringing Adaptivity and Robustness to DeFi Lending Markets

Decentralized Finance (DeFi) has revolutionized lending by replacing intermediaries with algorithm-driven liquidity pools. However, existing platforms like Aave and Compound rely on static interest rate curves and collateral requirements that struggle to adapt to rapid market changes, leading to inefficiencies in utilization and increased risks of liquidations. In this work, we propose a dynamic model of the lending market based on evolving demand and supply curves, alongside an adaptive interest rate controller that responds in real-time to shifting market conditions. Using a Recursive Least Squares algorithm, our controller tracks the external market and achieves stable utilization, while also controlling default and liquidation risk. We provide theoretical guarantees on the interest rate convergence and utilization stability of our algorithm. We establish bounds on the system's vulnerability to adversarial manipulation compared to static curves, while quantifying the trade-off between adaptivity and adversarial robustness. We propose two complementary approaches to mitigating adversarial manipulation: an algorithmic method that detects extreme demand and supply fluctuations and a market-based strategy that enhances elasticity, potentially via interest rate derivative markets. Our dynamic curve demand/supply model demonstrates a low best-fit error on Aave data, while our interest rate controller significantly outperforms static curve protocols in maintaining optimal utilization and minimizing liquidations.

cs.SI↗

arXiv2024

Thinking Fast and Slow: Data-Driven Adaptive DeFi Borrow-Lending Protocol

Decentralized finance (DeFi) borrowing and lending platforms are crucial to the decentralized economy, involving two main participants: lenders who provide assets for interest and borrowers who offer collateral exceeding their debt and pay interest. Collateral volatility necessitates over-collateralization to protect lenders and ensure competitive returns. Traditional DeFi platforms use a fixed interest rate curve based on the utilization rate (the fraction of available assets borrowed) and determine over-collateralization offline through simulations to manage risk. This method doesn't adapt well to dynamic market changes, such as price fluctuations and evolving user needs, often resulting in losses for lenders or borrowers. In this paper, we introduce an adaptive, data-driven protocol for DeFi borrowing and lending. Our approach includes a high-frequency controller that dynamically adjusts interest rates to maintain market stability and competitiveness with external markets. Unlike traditional protocols, which rely on user reactions and often adjust slowly, our controller uses a learning-based algorithm to quickly find optimal interest rates, reducing the opportunity cost for users during periods of misalignment with external rates. Additionally, we use a low-frequency planner that analyzes user behavior to set an optimal over-collateralization ratio, balancing risk reduction with profit maximization over the long term. This dual approach is essential for adaptive markets: the short-term component maintains market stability, preventing exploitation, while the long-term planner optimizes market parameters to enhance profitability and reduce risks. We provide theoretical guarantees on the convergence rates and adversarial robustness of the short-term component and the long-term effectiveness of our protocol. Empirical validation confirms our protocol's theoretical benefits.

cs.CE↗

arXiv2023

F3B: A Low-Overhead Blockchain Architecture with Per-Transaction Front-Running Protection

Front-running attacks, which benefit from advanced knowledge of pending transactions, have proliferated in the blockchain space since the emergence of decentralized finance. Front-running causes devastating losses to honest participants and continues to endanger the fairness of the ecosystem. We present Flash Freezing Flash Boys (F3B), a blockchain architecture that addresses front-running attacks by using threshold cryptography. In F3B, a user generates a symmetric key to encrypt their transaction, and once the underlying consensus layer has finalized the transaction, a decentralized secret-management committee reveals this key. F3B mitigates front-running attacks because, before the consensus group finalizes it, an adversary can no longer read the content of a transaction, thus preventing the adversary from benefiting from advanced knowledge of pending transactions. Unlike other mitigation systems, F3B properly ensures that all unfinalized transactions, even with significant delays, remain private by adopting per-transaction protection. Furthermore, F3B addresses front-running at the execution layer; thus, our solution is agnostic to the underlying consensus algorithm and compatible with existing smart contracts. We evaluated F3B on Ethereum with a modified execution layer and found only a negligible (0.026%) increase in transaction latency, specifically due to running threshold decryption with a 128-member secret-management committee after a transaction is finalized; this indicates that F3B is both practical and low-cost.

cs.CR↗

arXiv2023

Breaking Blockchain Rationality with Out-of-Band Collusion

Blockchain systems often rely on rationality assumptions for their security, expecting that nodes are motivated to maximize their profits. These systems thus design their protocols to incentivize nodes to execute the honest protocol but fail to consider out-of-band collusion. Existing works analyzing rationality assumptions are limited in their scope, either by focusing on a specific protocol or relying on non-existing financial instruments. We propose a general rational attack on rationality by leveraging an external channel that incentivizes nodes to collude against the honest protocol. Our approach involves an attacker creating an out-of-band bribery smart contract to motivate nodes to double-spend their transactions in exchange for shares in the attacker's profits. We provide a game theory model to prove that any rational node is incentivized to follow the malicious protocol. We discuss our approach to attacking the Bitcoin and Ethereum blockchains, demonstrating that irrational behavior can be rational in real-world blockchain systems when analyzing rationality in a larger ecosystem. We conclude that rational assumptions only appear to make the system more secure and offer a false sense of security under the flawed analysis.

cs.GT↗

arXiv2022

Online Admission Control and Rebalancing in Payment Channel Networks

Payment channel networks (PCNs) are a promising technology to improve the scalability of cryptocurrencies. PCNs, however, face the challenge that the frequent usage of certain routes may deplete channels in one direction, and hence prevent further transactions. In order to reap the full potential of PCNs, recharging and rebalancing mechanisms are required to provision channels, as well as an admission control logic to decide which transactions to reject in case capacity is insufficient. This paper presents a formal model of this optimisation problem. In particular, we consider an online algorithms perspective, where transactions arrive over time in an unpredictable manner. Our main contributions are competitive online algorithms which come with provable guarantees over time. We empirically evaluate our algorithms on randomly generated transactions to compare the average performance of our algorithms to our theoretical bounds. We also show how this model and approach differs from related problems in classic communication networks.

cs.DS↗