Etsy is a global marketplace where people across the world connect to make, buy and sell unique goods. Sellers at Etsy can promote their product listings via advertising campaigns similar to traditional sponsored search ads. Click-Through Rate (CTR) prediction is an integral part of online search advertising systems where it is utilized as an input to auctions which determine the final ranking of promoted listings to a particular user for each query. In this paper, we provide a holistic view of Etsy's promoted listings' CTR prediction system and propose an ensemble learning approach which is based on historical or behavioral signals for older listings as well as content-based features for new listings. We obtain representations from texts and images by utilizing state-of-the-art deep learning techniques and employ multimodal learning to combine these different signals. We compare the system to non-trivial baselines on a large-scale real world dataset from Etsy, demonstrating the effectiveness of the model and strong correlations between offline experiments and online performance. The paper is also the first technical overview to this kind of product in e-commerce context.
Embedding-based neural retrieval is a prevalent approach to address the semantic gap problem which often arises in product search on tail queries. In contrast, popular queries typically lack context and have a broad intent where additional context from users historical interaction can be helpful. In this paper, we share our novel approach to address both: the semantic gap problem followed by an end to end trained model for personalized semantic retrieval. We propose learning a unified embedding model incorporating graph, transformer and term-based embeddings end to end and share our design choices for optimal tradeoff between performance and efficiency. We share our learnings in feature engineering, hard negative sampling strategy, and application of transformer model, including a novel pre-training strategy and other tricks for improving search relevance and deploying such a model at industry scale. Our personalized retrieval model significantly improves the overall search experience, as measured by a 5.58% increase in search purchase rate and a 2.63% increase in site-wide conversion rate, aggregated across multiple A/B tests - on live traffic.
In this article, we present a general approach to personalizing ads through encoding and learning from variable-length sequences of recent user actions and diverse representations. To this end we introduce a three-component module called the adSformer diversifiable personalization module (ADPM) that learns a dynamic user representation. We illustrate the module's effectiveness and flexibility by personalizing the Click-Through Rate (CTR) and Post-Click Conversion Rate (PCCVR) models used in sponsored search. The first component of the ADPM, the adSformer encoder, includes a novel adSformer block which learns the most salient sequence signals. ADPM's second component enriches the learned signal through visual, multimodal, and other pretrained representations. Lastly, the third ADPM "learned on the fly" component further diversifies the signal encoded in the dynamic user representation. The ADPM-personalized CTR and PCCVR models, henceforth referred to as adSformer CTR and adSformer PCCVR, outperform the CTR and PCCVR production baselines by $+2.66\%$ and $+2.42\%$, respectively, in offline Area Under the Receiver Operating Characteristic Curve (ROC-AUC). Following the robust online
How do sellers on online marketplaces determine agreeable prices? This question is a theoretical concern for sociologists but a professional one for secondhand clothing resellers. Thousands of resellers across the United States purchase items from physical secondhand clothing sources and then resell them for a profit on sites such as Depop, Etsy, and Poshmark. They confront two pricing challenges: secondhand clothing items are aesthetic items of non-standard, uncertain quality, and online marketplaces offer limited explicit institutional support to back pricing claims. I analyze interviews and fieldwork to theorize how resellers price items for sale on online marketplaces. Resellers gain knowledge of secondhand community values and online marketplace technologies via immersion into offline (local reselling networks and secondhand sources) and online spaces (social media and the marketplaces themselves). Resellers selectively draw on these sources of pricing knowledge to deploy similar but varied pricing practices. These situated valuation practices reveal how resellers rely on reselling community structures and reflexively invoke pricing displays on marketplace interfaces to price secondhand clothing. These practices increase confidence in exchange as resellers can suitably justify the prices of material goods to online marketplace participants with varying levels of knowledge and experience.
This study is designed to remedy the tendency of existing studies to analyze online marketplaces as either sites of work or trading arenas. We argue that the theoretical notion of "social infrastructure" is particularly apt to offer a comprehensive framework that captures the unique intersection of work and trade in online marketplaces. We study the social infrastructure of an online marketplace: the institutions, conditions and forms, and the horizontal and vertical ties between actors that organize work and enable trading. The social infrastructure of online marketplaces deserves research attention because it represents an essential condition for economic activities. In our empirical section we focus on the online marketplace Etsy to illustrate our comprehensive theoretical framework and we identify a complex dynamic between the decided and emergent order of the online marketplace. We demonstrate that the attempt to superimpose order through the constitution of an online marketplace is challenged by sellers and buyers. We find that both dimensions, work and trade, provide actors with material and symbolic resources that inform their strategies and economic actions. The article suggests that "social infrastructure" is a concrete theoretical tool for analyzing online marketplaces that complements existing research on platforms and ecosystems.
Limited literature explores the relationship between economic impacts and pet ownership. Do people have more pets as a result of economic crises? In the current study, we answer this question by looking at the time series of pet ownership and children present in U.S. households from 2003 to 2018. We utilize a causal inference technique to compare the estimated real trend of pet ownership in American households against the scenario in which the 2007-2008 financial crisis would not have occurred. Our findings suggest that the financial crisis triggered households to own more pets, specifically dogs and cats.
Medication adherence is a problem of widespread concern in clinical care. Poor adherence is a particular problem for patients with chronic diseases requiring long-term medication because poor adherence can result in less successful treatment outcomes and even preventable deaths. Existing methods to collect information about patient adherence are resource-intensive or do not successfully detect low-adherers with high accuracy. Acknowledging that health measures recorded at clinic visits are more reliably recorded than a patient's adherence, we have developed an approach to infer medication adherence rates based on longitudinally recorded health measures that are likely impacted by time-varying adherence behaviors. Our framework permits the inclusion of baseline health characteristics and socio-demographic data. We employ a modular inferential approach. First, we fit a two-component model on a training set of patients who have detailed adherence data obtained from electronic medication monitoring. One model component predicts adherence behaviors only from baseline health and socio-demographic information, and the other predicts longitudinal health measures given the adherence and baseline health measures. Posterior draws of relevant model parameters are simulated from this model using Markov chain Monte Carlo methods. Second, we develop an approach to infer medication adherence from the time-varying health measures using a sequential Monte Carlo algorithm applied to a new set of patients for whom no adherence data are available. We apply and evaluate the method on a cohort of hypertensive patients, using baseline health comorbidities, socio-demographic measures, and blood pressure measured over time to infer patients' adherence to antihypertensive medication.
The COVID-19 pandemic simultaneously triggered a sudden, substantial increase in demand for items such as personal protection equipment and hospital ventilators whilst also disrupting the means of mass-production and international transport in established supply chains. Furthermore, under stay-at-home orders and with bricks-and-mortar retailers closed, consumers were also forced to adapt. Thus the pandemic offers a unique opportunity to study shifts in behaviour during disruption to industrialised manufacturing and economic contraction, in order to understand the role peer-to-peer production may play in a transition to long-term sustainability of production and consumption, or degrowth. Here, we analyse publicly-available datasets on internet search traffic and corporation financial returns to track the shifts in public interest and consumer behaviour over 2019 - 2020. We find a jump in interest in home-making and small-scale production at the beginning of the pandemic, as well as a substantial and sustained shift in consumer preference for peer-to-peer e-commerce platforms relative to more-established online vendors. In particular we present two case studies - the home-made facemasks supplied through Etsy, and the decentralised efforts of the 3D printer community - to assess the effectiveness of their responses to the pandemic. These patterns of behaviour are related to new modes of production in line with ecological economics and as such add capacity to a broader prefiguration of degrowth. We suggest an adoption of a new "fourth wave" of DIY culture defined by enhanced resilience and degrowth to continue to add capacity to a prefigurative politic of degrowth.
Background Clustering key clinical characteristics of participants in the Severe Asthma Research Program (SARP), a large, multicenter prospective observational study of patients with asthma and healthy controls, has led to the identification of novel asthma phenotypes. Purpose To determine whether quantitative CT (qCT) could help distinguish between clinical asthma phenotypes. Materials and Methods A retrospective cross-sectional analysis was conducted with the use of qCT images (maximal bronchodilation at total lung capacity [TLC], or inspiration, and functional residual capacity [FRC], or expiration) from the cluster phenotypes of SARP participants (cluster 1: minimal disease; cluster 2: mild, reversible; cluster 3: obese asthma; cluster 4: severe, reversible; cluster 5: severe, irreversible) enrolled between September 2001 and December 2015. Airway morphometry was performed along standard paths (RB1, RB4, RB10, LB1, and LB10). Corresponding voxels from TLC and FRC images were mapped with use of deformable image registration to characterize disease probability maps (DPMs) of functional small airway disease (fSAD), voxel-level volume changes (Jacobian), and isotropy (anisotropic deformation index [ADI]). The association between cluster assignment and qCT measures was evaluated using linear mixed models. Results A total of 455 participants were evaluated with cluster assignments and CT (mean age ± SD, 42.1 years ± 14.7; 270 women). Airway morphometry had limited ability to help discern between clusters. DPM fSAD was highest in cluster 5 (cluster 1 in SARP III: 19.0% ± 20.6; cluster 2: 18.9% ± 13.3; cluster 3: 24.9% ± 13.1; cluster 4: 24.1% ± 8.4; cluster 5: 38.8% ± 14.4; P < .001). Lower whole-lung Jacobian and ADI values were associated with greater cluster severity. Compared to cluster 1, cluster 5 lung expansion was 31% smaller (Jacobian in SARP III cohort: 2.31 ± 0.6 vs 1.61 ± 0.3, respectively, P < .001) and 34% more isotropic (ADI in SARP III cohort: 0.40 ± 0.1 vs 0.61 ± 0.2, P < .001). Within-lung Jacobian and ADI SDs decreased as severity worsened (Jacobian SD in SARP III cohort: 0.90 ± 0.4 for cluster 1; 0.79 ± 0.3 for cluster 2; 0.62 ± 0.2 for cluster 3; 0.63 ± 0.2 for cluster 4; and 0.41 ± 0.2 for cluster 5; P < .001). Conclusion Quantitative CT assessments of the degree and intraindividual regional variability of lung expansion distinguished between well-established clinical phenotypes among participants with asthma from the Severe Asthma Research Program study. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Verschakelen in this issue.
In response to the Coronavirus Disease 2019 (COVID-19) pandemic, current modeling supports the use of masks in community settings to reduce the transmission of SARS-CoV-2. However, concerns have been raised regarding the global shortage of medical grade masks and the limited evidence on the efficacy of fabric masks. This study used a standard mask testing method (ASTM F2101-14) and a model virus (bacteriophage MS2) to test the viral filtration efficiency (VFE) of fabric masks compared with commercially available disposable, surgical, and N95 masks. Five different types of fabric masks were purchased from the ecommerce website Etsy to represent a range of different fabric mask designs and materials currently available. One mask included a pocket for a filter; which was tested without a filter, with a dried baby wipe, and a section of a vacuum cleaner bag. A sixth fabric mask was also made according to the Victorian Department of Health and Human Services (DHHS) guidelines (Australia). Three masks of each type were tested. This study found that all the fabric masks had a VFE of at least 50% when tested against aerosols with an average size of 6.0 µm (VFE(6.0 µm)). The minimum VFE of fabric masks improved (to 63%) when the larger aerosols were excluded to give and average aerosol size of 2.6 µm (VFE(2.6 µm)), which better represents inhaled aerosols that can reach the lower respiratory system. The best performing fabric masks were the cotton mask with a section of vacuum cleaner bag (VFE(6.0 µm) = 99.5%, VFE(2.6 µm) = 98.8%) or a dried baby wipe (VFE(6.0 µm) = 98.5%, VFE(2.6 µm) = 97.6%) in the pocket designed for a disposable filter, the mask made using the Victorian DHHS design (VFE(6.0 µm) = 98.6%, VFE(2.6 µm) =99.1%) and one made from a layer of 100% hemp, a layer of poly membrane, and a layer of cheesecloth (VFE(6.0 µm) = 93.6%, VFE(2.6 µm) = 89.0%). The VFE of two surgical masks (VFE(6.0 µm) = 99.9% and 99.6%, VFE(2.6 µm) = 99.5% and 98.5%) and a N95 masks (VFE(6.0 µm) = 99.9%, VFE(2.6 µm) = 99.3%) were comparable to their advertised bacterial filtration efficacy. This research supports the use of fabric masks in the community to prevent the spread of SARS-CoV-2; however, future research is needed to explore the optimum design in ensuring proper fit. There is also a need for mass education campaigns to disseminate this information, along with guidelines around the proper usage and washing of fabric masks.
Recent data suggest novel functional roles for cerebellar involvement in a number of neurologic diseases. Function of cerebellar neurons is known to be modulated by norepinephrine and adrenergic receptors. The distribution of adrenergic receptor subtypes has been described in experimental animals, but corroboration of such studies in the human cerebellum, necessary for drug treatment, is still lacking. In the present work we studied cell-specific localizations of alpha1 adrenergic receptor subtype mRNA (alpha 1a, alpha 1b, alpha 1d), and alpha2 adrenergic receptor subtype mRNA (alpha 2a, alpha 2b, alpha 2c) by in situ hybridization on cryostat sections of human cerebellum (cortical layers and dentate nucleus). We observed unique neuron-specific alpha1 adrenergic receptor and alpha2 adrenergic receptor subtype distribution in human cerebellum. The cerebellar cortex expresses mRNA encoding all six alpha adrenergic receptor subtypes, whereas dentate nucleus neurons express all subtype mRNAs, except alpha 2a adrenergic receptor mRNA. All Purkinje cells label strongly for alpha 2a and alpha 2b adrenergic receptor mRNA. Additionally, Purkinje cells of the anterior lobe vermis (lobules I to V) and uvula/tonsil (lobules IX/HIX) express alpha 1a and alpha 2c subtypes, and Purkinje cells in the ansiform lobule (lobule HVII) and uvula/tonsil express alpha 1b and alpha 2c adrenergic receptor subtypes. Basket cells show a strong signal for alpha 1a, moderate signal for alpha 2a and light label for alpha 2b adrenergic receptor mRNA. In stellate cells, besides a strong label of alpha 2a adrenergic receptor mRNA in all and moderate label of alpha 2b message in select stellate cells, the inner stellate cells are also moderately positive for alpha 1b adrenergic receptor mRNA. Granule and Golgi cells express high levels of alpha 2a and alpha 2b adrenergic receptor mRNAs. These data contribute new information regarding specific location of adrenergic receptor subtypes in human cerebellar neurons. We discuss our observations in terms of possible modulatory roles of adrenergic receptor subtypes in cerebellar neurons responding to sensory and autonomic input signals, and review species differences in cerebellar adrenergic receptor expression.
AI agents are becoming active decision-makers on the Internet. As they make decisions in the same environments as humans, the environments themselves can change to influence them. We call this $\textit{mecha-nudging}$: changes to how choices are presented that systematically influence AI agents without materially degrading the decision environment for humans. To measure this phenomenon, we combine two frameworks -- Bayesian persuasion from economics and $\mathcal{V}$-usable information from computer science -- to get a common unit (bits) for quantifying how environments change across a wide range of interventions, contexts, and models. We apply this framework to over six million Etsy listings and find that, after ChatGPT's release, listings contain significantly more machine-usable information for predicting agent curation decisions, increasing by 0.143 bits out of a maximum possible increase of 0.355. This shift is robust across prompts, token choices, labeling models, and fine-tuning architectures; absent in a regulated-text placebo; and far larger than the effect of generic LLM rewriting. In contrast, a human study finds little to no change in human-usable information. Our resul
This letter evaluates the performance of the slotted Aloha protocol defined by the European Telecommunication Standard Institute (ETSI) SmartBAN specification, under saturation conditions. For this purpose, we develop a two-dimensional Discrete Time Markov Chain (DTMC) to model the operational details of the protocol and assess its performance in terms of saturation throughput and average end-to-end delay. The accuracy of the proposed model is validated by means of simulation which reveals a very good match among theoretical and simulation results. The model can be used for protocol performance prediction and optimization purposes.
As V2X (Vehicle-to-Everything) technology becomes increasingly prevalent, the security of V2X networks has garnered growing attention worldwide. In North America, the IEEE 1609 series standards are primarily used, while Europe adopts the ETSI series standards, and China has also established its industry standard, YD/T 3957-2021, among others. Although these standards share some commonalities, they also exhibit differences. To achieve compatibility across these standards, analyzing their similarities and differences is a crucial issue. Therefore, this study focuses on analyzing the three major standards mentioned above, discussing aspects such as certificate formats, signed message formats, and certificate request processes. Additionally, this research evaluates the efficiency of different cryptographic methods, including NIST P-256 and SM2-256, SHA-256 and SM3-256, as well as AES-128 and SM4-128. Finally, the study implements these three major standards on V2X devices and compares the efficiency of message signing and signature verification in V2X systems, providing a reference for the development of a secure certificate management system for V2X networks.
This report presents a prototype VPN that uses QKD-derived keys delivered through the ETSI GS QKD 014 API. The VPN encrypts IP traffic with AES-256-GCM, transports ETSI key identifiers in-band, and retrieves matching keys from local KMEs. After validation with a controlled KME simulator, the system was tested on two Jetson Xavier NX devices connected to a LuxQuanta NOVA QKD platform. The experiment successfully transmitted bidirectional real-time audio and video traffic through the VPN for eight continuous hours, demonstrating the feasibility of integrating classical VPN applications with QKD infrastructure through a standardized key-delivery interface.
Efficient maneuver coordination in dense V2X environments requires accurate short-term prediction while maintaining low communication and computational overhead. Current European Telecommunications Standards Institute (ETSI)-compliant approaches rely on intention detection and trajectory vector transmission, which scale poorly with neighborhood size and prediction horizon. This paper revisits maneuver coordination from an intention sharing perspective and investigates geometric encodings that enable scalable communication. First, we analyze three ETSI-compliant encodings, trajectory vectors, N-polygons, and uncertainty ellipses, through complexity analysis and simulation-based CPU measurements. Results show that uncertainty ellipses reduce computational complexity by an order of magnitude compared with trajectory vectors while maintaining a constant message size. Building on this, an Extended Kalman Filter is used to generate short-horizon predictions, which are encoded as uncertainty ellipses to represent the intended maneuver. The prediction pipeline is evaluated using real-world GNSS trajectories collected from cyclist maneuvers on a controlled test track, demonstrating that the
The automotive sector is following a revolutionary path from vehicles controlled by humans to vehicles that will be fully automated, fully connected, and ultimately fully cooperative. Along this road, new cooperative algorithms and protocols will be designed and field tested, which represents a great challenge in terms of costs. In this context, in particular, moving from simulations to practical experiments requires huge investments that are not always affordable and may become a barrier in some cases. To solve this issue and provide the community with an intermediate step, we here propose the use of 1:10 scaled cooperative, autonomous, and connected mini-cars. The mini-car is equipped with a Jetson Orin board running the open Robot Operating System 2 (ROS2), sensors for autonomous operations, and a Raspberry Pi board for connectivity mounting the open source Open Stack for Car (OScar). A key aspect of the proposal is the use of OScar, which implements a full ETSI cooperative-intelligent transport systems (C-ITS) compliant stack. The feasibility and potential of the proposed platform is here demonstrated through the implementation of a case study where the Day-1 intersection colli
Industrial standards and normative documents exhibit intricate hierarchical structures, domain-specific lexicons, and extensive cross-referential dependencies, which making it challenging to process them directly by Large Language Models (LLMs). While Retrieval-Augmented Generation (RAG) provides a computationally efficient alternative to LLM fine-tuning, standard "vanilla" vector-based retrieval may fail to capture the latent structural and relational features intrinsic in normative documents. With the objective of shedding light on the most promising technique for building high-performance RAG solutions for normative, standards, and regulatory documents, this paper investigates the efficacy of Graph RAG architectures, which represent information as interconnected nodes, thus moving from simple semantic similarity toward a more robust, relation-aware retrieval mechanism. Despite the promise of graph-based techniques, there is currently a lack of empirical evidence as to which is the optimal indexing strategy for technical standards. Therefore, to help solve this knowledge gap, we propose a specialized RAG methodology tailored to the unique structure and lexical characteristics of
The rapid growth of e-commerce requires robust multimodal representations that capture diverse signals from user-generated listings. Existing vision-language models (VLMs) typically align titles with primary images, i.e., single-view, but overlook non-primary images and auxiliary textual views that provide critical semantics in open marketplaces such as Etsy or Poshmark. To this end, we propose a framework that unifies multimodal and multi-view learning through Factorized Transport, a lightweight approximation of optimal transport, designed for scalability and deployment efficiency. During training, the method emphasizes primary views while stochastically sampling auxiliary ones, reducing training cost from quadratic in the number of views to constant per item. At inference, all views are fused into a single cached embedding, preserving the efficiency of two-tower retrieval with no additional online overhead. On an industrial dataset of 1M product listings and 0.3M interactions, our approach delivers consistent improvements in cross-view and query-to-item retrieval, achieving up to +7.9% Recall@500 over strong multimodal baselines. Overall, our framework bridges scalability with op
Integrated Sensing and Communications (ISAC) is defined as one of six usage scenarios in the ITU-R International Mobile Telecommunications (IMT) 2030 framework for 6G. ISAC is envisioned to introduce the sensing capability into the cellular network, where sensing may be obtained using the cellular radio frequency (RF) signals with or without additional auxiliary sensors. To enable ISAC, specification bodies such as European Telecommunications Standards Institute (ETSI) and Third Generation Partnership Project (3GPP) have already started to look into detailed ISAC use cases, their requirements, and the channel models and evaluation methodologies that are necessary to design and evaluate ISAC performance. With focus on the channel model, the current communication-centric channel models like those specified in 3GPP technical report (TR) 38.901 do not cover the RF signals interactions between the transmitter, target object, receiver and their surrounding environment. To bridge this gap, 3GPP has been looking into the basic changes that are necessary to make to their TR38.901 channel model with focus on selected use cases from the 3GPP SA1 5G-Advanced feasibility study. In parallel, ETS