Search

The detection of deepfake speech has become increasingly challenging with the rapid evolution of deepfake technologies. In this paper, we propose a hybrid architecture for deepfake speech detection, combining a self-supervised learning framework for feature extraction with a classifier head to form an end-to-end model. Our approach incorporates both audio-level and feature-level augmentation techniques. Specifically, we introduce and analyze various masking strategies for augmenting raw audio spectrograms and for enhancing feature representations during training. We incorporate compression augmentations during the pretraining phase of the feature extractor to address the limitations of small, single-language datasets. We evaluate the model on the ASVSpoof5 (ASVSpoof 2024) challenge, achieving state-of-the-art results in Track 1 under closed conditions with an Equal Error Rate of 4.37%. By employing different pretrained feature extractors, the model achieves an enhanced EER of 3.39%. Our model demonstrates robust performance against unseen deepfake attacks and exhibits strong generalization across different codecs.

A contract is an economic tool used by a principal to incentivize one or more agents to exert effort on her behalf, by defining payments based on observable performance measures. A key challenge addressed by contracts -- known in economics as moral hazard -- is that, absent a properly set up contract, agents might engage in actions that are not in the principal's best interest. Another common feature of contracts is limited liability, which means that payments can go only from the principal -- who has the deep pocket -- to the agents. With classic applications of contract theory moving online, growing in scale, and becoming more data-driven, tools from contract theory become increasingly important for incentive-aware algorithm design. At the same time, algorithm design offers a whole new toolbox for reasoning about contracts, ranging from additional tools for studying the tradeoff between simple and optimal contracts, through a language for discussing the computational complexity of contracts in combinatorial settings, to a formalism for analyzing data-driven contracts. This survey aims to provide a computer science-friendly introduction to the basic concepts of contract theory. We give an overview of the emerging field of "algorithmic contract theory" and highlight work that showcases the potential for interaction between the two areas. We also discuss avenues for future research.

Energy efficient, safe and reliable Li-ion batteries (LIBs) are required for a wide range of applications. Charging capabilities of thick electrodes still holding their stored high-energy is a most desirable characteristic in future advanced LIBs. The introduction of ultra-thick graphite anode meets limitations in internal electrode transport properties, leading to Li-ion gradients with detrimental consequences for battery cell performance and lifetime. Yet, there is a lack of experimental tools capable of providing a complete view of local processes and evolving gradients within such thick electrodes. Here, we introduce a multi-modal operando measurement approach, enabling quantitative spatio-temporal observations of Li concentrations and intercalation phases in ultra-thick, graphite electrodes. Neutron imaging and diffraction concurrently provide correlated information from the macroscopic scale of the cell and electrode down to the crystallographic scale portraying the intercalation and deintercalation processes. In particular, the evolving formation of the solid electrolyte interphase (SEI), observation of gradients in total lithium content, as well as in the formation of ordered LixC6 phases and trapped lithium have been mapped throughout the first charge-discharge cycle of the cell. Different lithiation stages co-exist during charging and discharging of an ultra-thick composite graphite-based electrode; delayed lithiation and delithiation processes are observed at the central region of the electrode, while the SEI formation, potential plating and dead lithium are predominantly found closer to the interface with the separator. The study furthermore emphasizes the potential of the method to study Li ion diffusion and the kinetics of lithiation phase formation in advanced ultra-thick electrodes.

Shared control problems involve a robot learning to collaborate with a human. When learning a shared control policy, short communication between the agents can often significantly reduce running times and improve the system's accuracy. We extend the shared control problem to include the ability to directly query a cooperating agent. We consider two types of potential responses to a query, namely oracles: one that can provide the learner with the best action they should take, even when that action might be myopically wrong, and one with a bounded knowledge limited to its part of the system. Given this additional information channel, this work further presents three heuristics for choosing when to query: reinforcement learning-based, utility-based, and entropy-based. These heuristics aim to reduce a system's overall learning cost. Empirical results on two environments show the benefits of querying to learn a better control policy and the tradeoffs between the proposed heuristics.

The rapidly evolving field of digital oncopathology faces significant challenges, including the need to address diverse and complex clinical questions, often involving rare conditions, with limited availability of labeled data. These limitations hinder the development of robust AI-driven tools in the biomedical space, where accuracy in probabilistic determinations is of utmost importance. To address this, digital pathology foundation models have begun to emerge, typically developed with the size and diversity of the pre-training dataset and model parameters in mind. Here, we present CanvOI, a ViT-g/10-based foundation model designed to enhance the capabilities of digital pathology by addressing these challenges through a different approach. Considering the unique nature of oncologic histopathological images and the requirements from the embeddings to provide meaningful representations for Multiple Instance Learning (MIL) downstream models, we chose to modify the input image characteristics. By introducing larger tile sizes (380 x 380 pixels) and smaller patch sizes (10 x 10 pixels), we were able to optimize the model's performance, pushing computational resources in a new direction and achieving state-of-the-art performance on cancer-related benchmarks. CanvOI demonstrated a 1.5-7.4% improvement in averaged AUC compared to other leading foundation models built for digital pathology. Moreover, our results demonstrate that CanvOI significantly outperformed the other models, with the performance gap widening substantially when trained on just 10% of the initial cohort. This work highlights an alternative approach that, if integrated with traditional development approaches, has the potential to advance Oncology Intelligence (OI), overcome some of the current barriers and ultimately improve the clinical outcome of cancer patients., Comment: 12 pages, 5 figures

Dental disease is a prevalent chronic condition associated with substantial financial burden, personal suffering, and increased risk of systemic diseases. Despite widespread recommendations for twice-daily tooth brushing, adherence to recommended oral self-care behaviors remains sub-optimal due to factors such as forgetfulness and disengagement. To address this, we developed Oralytics, a mHealth intervention system designed to complement clinician-delivered preventative care for marginalized individuals at risk for dental disease. Oralytics incorporates an online reinforcement learning algorithm to determine optimal times to deliver intervention prompts that encourage oral self-care behaviors. We have deployed Oralytics in a registered clinical trial. The deployment required careful design to manage challenges specific to the clinical trials setting in the U.S. In this paper, we (1) highlight key design decisions of the RL algorithm that address these challenges and (2) conduct a re-sampling analysis to evaluate algorithm design decisions. A second phase (randomized control trial) of Oralytics is planned to start in spring 2025.

Online AI decision-making algorithms are increasingly used by digital interventions to dynamically personalize treatment to individuals. These algorithms determine, in real-time, the delivery of treatment based on accruing data. The objective of this paper is to provide guidelines for enabling effective monitoring of online decision-making algorithms with the goal of (1) safeguarding individuals and (2) ensuring data quality. We elucidate guidelines and discuss our experience in monitoring online decision-making algorithms in two digital intervention clinical trials (Oralytics and MiWaves). Our guidelines include (1) developing fallback methods, pre-specified procedures executed when an issue occurs, and (2) identifying potential issues categorizing them by severity (red, yellow, and green). Across both trials, the monitoring systems detected real-time issues such as out-of-memory issues, database timeout, and failed communication with an external source. Fallback methods prevented participants from not receiving any treatment during the trial and also prevented the use of incorrect data in statistical analyses. These trials provide case studies for how health scientists can build monitoring systems for their digital intervention. Without these algorithm monitoring systems, critical issues would have gone undetected and unresolved. Instead, these monitoring systems safeguarded participants and ensured the quality of the resulting data for updating the intervention and facilitating scientific discovery. These monitoring guidelines and findings give digital intervention teams the confidence to include online decision-making algorithms in digital interventions.

The escalating prevalence of cannabis use poses a significant public health challenge globally. In the U.S., cannabis use is more prevalent among emerging adults (EAs) (ages 18-25) than any other age group, with legalization in the multiple states contributing to a public perception that cannabis is less risky than in prior decades. To address this growing concern, we developed MiWaves, a reinforcement learning (RL) algorithm designed to optimize the delivery of personalized intervention prompts to reduce cannabis use among EAs. MiWaves leverages domain expertise and prior data to tailor the likelihood of delivery of intervention messages. This paper presents a comprehensive overview of the algorithm's design, including key decisions and experimental outcomes. The finalized MiWaves RL algorithm was deployed in a clinical trial from March to May 2024., Comment: arXiv admin note: substantial text overlap with arXiv:2402.17739

We present Jamba-1.5, new instruction-tuned large language models based on our Jamba architecture. Jamba is a hybrid Transformer-Mamba mixture of experts architecture, providing high throughput and low memory usage across context lengths, while retaining the same or better quality as Transformer models. We release two model sizes: Jamba-1.5-Large, with 94B active parameters, and Jamba-1.5-Mini, with 12B active parameters. Both models are fine-tuned for a variety of conversational and instruction-following capabilties, and have an effective context length of 256K tokens, the largest amongst open-weight models. To support cost-effective inference, we introduce ExpertsInt8, a novel quantization technique that allows fitting Jamba-1.5-Large on a machine with 8 80GB GPUs when processing 256K-token contexts without loss of quality. When evaluated on a battery of academic and chatbot benchmarks, Jamba-1.5 models achieve excellent results while providing high throughput and outperforming other open-weight models on long-context benchmarks. The model weights for both sizes are publicly available under the Jamba Open Model License and we release ExpertsInt8 as open source., Comment: Webpage: https://www.ai21.com/jamba

The increasing deployment of AI is shaping the future landscape of the internet, which is set to become an integrated ecosystem of AI agents. Orchestrating the interaction among AI agents necessitates decentralized, self-sustaining mechanisms that harmonize the tension between individual interests and social welfare. In this paper we tackle this challenge by synergizing reinforcement learning with principal-agent theory from economics. Taken separately, the former allows unrealistic freedom of intervention, while the latter struggles to scale in sequential settings. Combining them achieves the best of both worlds. We propose a framework where a principal guides an agent in a Markov Decision Process (MDP) using a series of contracts, which specify payments by the principal based on observable outcomes of the agent's actions. We present and analyze a meta-algorithm that iteratively optimizes the policies of the principal and agent, showing its equivalence to a contraction operator on the principal's Q-function, and its convergence to subgame-perfect equilibrium. We then scale our algorithm with deep Q-learning and analyze its convergence in the presence of approximation error, both theoretically and through experiments with randomly generated binary game-trees. Extending our framework to multiple agents, we apply our methodology to the combinatorial Coin Game. Addressing this multi-agent sequential social dilemma is a promising first step toward scaling our approach to more complex, real-world instances.

In recent years, notable advancements have been made in the domain of visual document understanding, with the prevailing architecture comprising a cascade of vision and language models. The text component can either be extracted explicitly with the use of external OCR models in OCR-based approaches, or alternatively, the vision model can be endowed with reading capabilities in OCR-free approaches. Typically, the queries to the model are input exclusively to the language component, necessitating the visual features to encompass the entire document. In this paper, we present VisFocus, an OCR-free method designed to better exploit the vision encoder's capacity by coupling it directly with the language prompt. To do so, we replace the down-sampling layers with layers that receive the input prompt and allow highlighting relevant parts of the document, while disregarding others. We pair the architecture enhancements with a novel pre-training task, using language masking on a snippet of the document text fed to the visual encoder in place of the prompt, to empower the model with focusing capabilities. Consequently, VisFocus learns to allocate its attention to text patches pertinent to the provided prompt. Our experiments demonstrate that this prompt-guided visual encoding approach significantly improves performance, achieving state-of-the-art results on various benchmarks., Comment: 32 pages, 18 figures

Deep Active Learning (AL) techniques can be effective in reducing annotation costs for training deep models. However, their effectiveness in low- and high-budget scenarios seems to require different strategies, and achieving optimal results across varying budget scenarios remains a challenge. In this study, we introduce Dynamic Coverage & Margin mix (DCoM), a novel active learning approach designed to bridge this gap. Unlike existing strategies, DCoM dynamically adjusts its strategy, considering the competence of the current model. Through theoretical analysis and empirical evaluations on diverse datasets, including challenging computer vision tasks, we demonstrate DCoM's ability to overcome the cold start problem and consistently improve results across different budgetary constraints. Thus DCoM achieves state-of-the-art performance in both low- and high-budget regimes.

Dental disease is still one of the most common chronic diseases in the United States. While dental disease is preventable through healthy oral self-care behaviors (OSCB), this basic behavior is not consistently practiced. We have developed Oralytics, an online, reinforcement learning (RL) algorithm that optimizes the delivery of personalized intervention prompts to improve OSCB. In this paper, we offer a full overview of algorithm design decisions made using prior data, domain expertise, and experiments in a simulation test bed. The finalized RL algorithm was deployed in the Oralytics clinical trial, conducted from fall 2023 to summer 2024.

While the success of large language models (LLMs) increases demand for machine-generated text, current pay-per-token pricing schemes create a misalignment of incentives known in economics as moral hazard: Text-generating agents have strong incentive to cut costs by preferring a cheaper model over the cutting-edge one, and this can be done "behind the scenes" since the agent performs inference internally. In this work, we approach this issue from an economic perspective, by proposing a pay-for-performance, contract-based framework for incentivizing quality. We study a principal-agent game where the agent generates text using costly inference, and the contract determines the principal's payment for the text according to an automated quality evaluation. Since standard contract theory is inapplicable when internal inference costs are unknown, we introduce cost-robust contracts. As our main theoretical contribution, we characterize optimal cost-robust contracts through a direct correspondence to optimal composite hypothesis tests from statistics, generalizing a result of Saig et al. (NeurIPS'23). We evaluate our framework empirically by deriving contracts for a range of objectives and LLM evaluation benchmarks, and find that cost-robust contracts sacrifice only a marginal increase in objective value compared to their cost-aware counterparts., Comment: Comments are welcome

Auctions are key for maximizing sellers' revenue and ensuring truthful bidding among buyers. Recently, an approach known as differentiable economics based on deep learning shows promise in learning optimal auction mechanisms for multiple items and participants. However, this approach has no guarantee of strategy-proofness at test time. Strategy-proofness is crucial as it ensures that buyers are incentivized to bid their true valuations, leading to optimal and fair auction outcomes without the risk of manipulation. Building on conformal prediction, we introduce a novel approach to achieve strategy-proofness with rigorous statistical guarantees. The key novelties of our method are: (i) the formulation of a regret prediction model, used to quantify at test time violations of strategy-proofness; and (ii) an auction acceptance rule that leverages the predicted regret to ensure that for a new auction, the data-driven mechanism meets the strategy-proofness requirement with high probability (e.g., 99\%). Numerical experiments demonstrate the necessity for rigorous guarantees, the validity of our theoretical results, and the applicability of our proposed method.

The availability of mobile health (mHealth) technology has enabled increased collection of intensive longitudinal data (ILD). ILD have potential to capture rapid fluctuations in outcomes that may be associated with changes in the risk of an event. However, existing methods for jointly modeling longitudinal and event-time outcomes are not well-equipped to handle ILD due to the high computational cost. We propose a joint longitudinal and time-to-event model suitable for analyzing ILD. In this model, we summarize a multivariate longitudinal outcome as a smaller number of time-varying latent factors. These latent factors, which are modeled using an Ornstein-Uhlenbeck stochastic process, capture the risk of a time-to-event outcome in a parametric hazard model. We take a Bayesian approach to fit our joint model and conduct simulations to assess its performance. We use it to analyze data from an mHealth study of smoking cessation. We summarize the longitudinal self-reported intensity of nine emotions as the psychological states of positive and negative affect. These time-varying latent states capture the risk of the first smoking lapse after attempted quit. Understanding factors associated with smoking lapse is of keen interest to smoking cessation researchers., Comment: Main text is 32 pages with 6 figures. Supplementary material is 21 pages

Existing statistical methods for the analysis of micro-randomized trials (MRTs) are designed to estimate causal excursion effects using data from a single MRT. In practice, however, researchers can often find previous MRTs that employ similar interventions. In this paper, we develop data integration methods that capitalize on this additional information, leading to statistical efficiency gains. To further increase efficiency, we demonstrate how to combine these approaches according to a generalization of multivariate precision weighting that allows for correlation between estimates, and we show that the resulting meta-estimator possesses an asymptotic optimality property. We illustrate our methods in simulation and in a case study involving two MRTs in the area of smoking cessation.

Algorithms with predictions have attracted much attention in the last years across various domains, including variants of facility location, as a way to surpass traditional worst-case analyses. We study the $k$-facility location mechanism design problem, where the $n$ agents are strategic and might misreport their location. Unlike previous models, where predictions are for the $k$ optimal facility locations, we receive $n$ predictions for the locations of each of the agents. However, these predictions are only "mostly" and "approximately" correct (or MAC for short) -- i.e., some $\delta$-fraction of the predicted locations are allowed to be arbitrarily incorrect, and the remainder of the predictions are allowed to be correct up to an $\varepsilon$-error. We make no assumption on the independence of the errors. Can such predictions allow us to beat the current best bounds for strategyproof facility location? We show that the $1$-median (geometric median) of a set of points is naturally robust under corruptions, which leads to an algorithm for single-facility location with MAC predictions. We extend the robustness result to a "balanced" variant of the $k$ facilities case. Without balancedness, we show that robustness completely breaks down, even for the setting of $k=2$ facilities on a line. For this "unbalanced" setting, we devise a truthful random mechanism that outperforms the best known result of Lu et al. [2010], which does not use predictions. En route, we introduce the problem of "second" facility location (when the first facility's location is already fixed). Our findings on the robustness of the $1$-median and more generally $k$-medians may be of independent interest, as quantitative versions of classic breakdown-point results in robust statistics.

The escalating prevalence of cannabis use, and associated cannabis-use disorder (CUD), poses a significant public health challenge globally. With a notably wide treatment gap, especially among emerging adults (EAs; ages 18-25), addressing cannabis use and CUD remains a pivotal objective within the 2030 United Nations Agenda for Sustainable Development Goals (SDG). In this work, we develop an online reinforcement learning (RL) algorithm called reBandit which will be utilized in a mobile health study to deliver personalized mobile health interventions aimed at reducing cannabis use among EAs. reBandit utilizes random effects and informative Bayesian priors to learn quickly and efficiently in noisy mobile health environments. Moreover, reBandit employs Empirical Bayes and optimization techniques to autonomously update its hyper-parameters online. To evaluate the performance of our algorithm, we construct a simulation testbed using data from a prior study, and compare against commonly used algorithms in mobile health studies. We show that reBandit performs equally well or better than all the baseline algorithms, and the performance gap widens as population heterogeneity increases in the simulation environment, proving its adeptness to adapt to diverse population of study participants.

Online reinforcement learning (RL) algorithms offer great potential for personalizing treatment for participants in clinical trials. However, deploying an online, autonomous algorithm in the high-stakes healthcare setting makes quality control and data quality especially difficult to achieve. This paper proposes algorithm fidelity as a critical requirement for deploying online RL algorithms in clinical trials. It emphasizes the responsibility of the algorithm to (1) safeguard participants and (2) preserve the scientific utility of the data for post-trial analyses. We also present a framework for pre-deployment planning and real-time monitoring to help algorithm developers and clinical researchers ensure algorithm fidelity. To illustrate our framework's practical application, we present real-world examples from the Oralytics clinical trial. Since Spring 2023, this trial successfully deployed an autonomous, online RL algorithm to personalize behavioral interventions for participants at risk for dental disease.

Many real-life contractual relations differ completely from the clean, static model at the heart of principal-agent theory. Typically, they involve repeated strategic interactions of the principal and agent, taking place under uncertainty and over time. While appealing in theory, players seldom use complex dynamic strategies in practice, often preferring to circumvent complexity and approach uncertainty through learning. We initiate the study of repeated contracts with a learning agent, focusing on agents who achieve no-regret outcomes. Optimizing against a no-regret agent is a known open problem in general games; we achieve an optimal solution to this problem for a canonical contract setting, in which the agent's choice among multiple actions leads to success/failure. The solution has a surprisingly simple structure: for some $\alpha > 0$, initially offer the agent a linear contract with scalar $\alpha$, then switch to offering a linear contract with scalar $0$. This switch causes the agent to ``free-fall'' through their action space and during this time provides the principal with non-zero reward at zero cost. Despite apparent exploitation of the agent, this dynamic contract can leave \emph{both} players better off compared to the best static contract. Our results generalize beyond success/failure, to arbitrary non-linear contracts which the principal rescales dynamically. Finally, we quantify the dependence of our results on knowledge of the time horizon, and are the first to address this consideration in the study of strategizing against learning agents.

Data storage in DNA is developing as a possible solution for archival digital data. Recently, to further increase the potential capacity of DNA-based data storage systems, the combinatorial composite DNA synthesis method was suggested. This approach extends the DNA alphabet by harnessing short DNA fragment reagents, known as shortmers. The shortmers are building blocks of the alphabet symbols, consisting of a fixed number of shortmers. Thus, when information is read, it is possible that one of the shortmers that forms part of the composition of a symbol is missing and therefore the symbol cannot be determined. In this paper, we model this type of error as a type of asymmetric error and propose code constructions that can correct such errors in this setup. We also provide a lower bound on the redundancy of such error-correcting codes and give an explicit encoder and decoder pair for our construction. Our suggested error model is also supported by an analysis of data from actual experiments that produced DNA according to the combinatorial scheme. Lastly, we also provide a statistical evaluation of the probability of observing such error events, as a function of read depth.

The Tigray War was an armed conflict that took place primarily in the Tigray region of northern Ethiopia from November 3, 2020 to November 2, 2022. Given the importance of agriculture in Tigray to livelihoods and food security, determining the impact of the war on cultivated area is critical. However, quantifying this impact was difficult due to restricted movement within and into the region and conflict-driven insecurity and blockages. Using satellite imagery and statistical area estimation techniques, we assessed changes in crop cultivation area in Tigray before and during the war. Our findings show that cultivated area was largely stable between 2020-2021 despite the widespread impacts of the war. We estimated $1,132,000\pm133,000$ hectares of cultivation in pre-war 2020 compared to $1,217,000 \pm 132,000$ hectares in wartime 2021. Comparing changes inside and outside of a 5 km buffer around conflict events, we found a slightly higher upper confidence limit of cropland loss within the buffer (0-3%) compared to outside the buffer (0-1%). Our results support other reports that despite widespread war-related disruptions, Tigrayan farmers were largely able to sustain cultivation. Our study demonstrates the capability of remote sensing combined with machine learning and statistical techniques to provide timely, transparent area estimates for monitoring food security in regions inaccessible due to conflict.