Hasil untuk "Revenue. Taxation. Internal revenue"

F. Chaloupka, A. Yurekli, G. Fong

Samuel Yeh, Sharon Li

Retrieval-augmented generation (RAG) enhances large language models (LLMs) by conditioning generation on retrieved external documents, but the effect of retrieved context is often non-trivial. In realistic retrieval settings, the retrieved document set often contains a mixture of documents that vary in relevance and usefulness. While prior work has largely examined these phenomena through output behavior, little is known about how retrieved context shapes the internal representations that mediate information integration in RAG. In this work, we study RAG through the lens of latent representations. We systematically analyze how different types of retrieved documents affect the hidden states of LLMs, and how these internal representation shifts relate to downstream generation behavior. Across four question-answering datasets and three LLMs, we analyze internal representations under controlled single- and multi-document settings. Our results reveal how context relevancy and layer-wise processing influence internal representations, providing explanations on LLMs output behaviors and insights for RAG system design.

Xinyu Wang, Hanwei Wu, Jingwei Song et al.

Large language model (LLM) agents learn by interacting with environments, but long-horizon training remains fundamentally bottlenecked by sparse and delayed rewards. Existing methods typically address this challenge through post-hoc credit assignment or external reward models, which provide limited guidance at inference time and often separate reward improvement from policy improvement. We propose Self-Guide, a self-generated internal reward for language agents that supports both inference-time guidance and training-time supervision. Specifically, the agent uses Self-Guide as a short self-guidance signal to steer the next action during inference, and converts the same signal into step-level internal reward for denser policy optimization during training. This creates a co-evolving loop: better policy produces better guidance, and better guidance further improves policy as internal reward. Across three agent benchmarks, inference-time self-guidance already yields clear gains, while jointly evolving policy and internal reward with GRPO brings further improvements (8\%) over baselines trained solely with environment reward. Overall, our results suggest that language agents can improve not only by collecting more experience, but also by learning to generate and refine their own internal reward during acting and learning.

Alexandra J. Simpson, Jacqueline M. McSweeney, James A. Lerczak et al.

A combined radar remote sensing and in situ data set is used to track packets of nonlinear internal waves as they propagate and shoal across the inner shelf (40m - 9m). The dataset consists of high space-time resolution (5m, 2min) radar image time series collected over a 10km radial footprint, with over a dozen synchronous and co-located moorings measuring temperature, salinity, and velocity throughout the water column. The internal bores and higher-frequency internal waves that make up the internal tide are tracked in the radar image time series and, therefore, provide continuous cross-shore speed and angle estimates as the waves propagate across the inner shelf. We compare radar-estimated speeds to those estimated from moorings and confirm a cross-shore shoaling profile that deviates from linear theory. We additionally use combined remote sensing and in situ data to perform a detailed analysis on four consecutive internal tides. These analyses reveal intra-packet speed variability, tide-tide influences, reflected internal waves, and an instance of internal wave polarity reversal observed in the radar and moorings.

Yusuke Morimoto

This paper introduces a new algorithmic execution model that integrates interbank limit and market orders with internal liquidity generated through market making. Based on the Cartea et al.\cite{cartea2015algorithmic} framework, we incorporate market impact in interbank orders while excluding it for internal market-making transactions. Our model aims to optimize the balance between interbank and internal liquidity, reducing market impact and improving execution efficiency.

E. V. Semenova, D. V. Sidorov

Objective: to critically analyze the international experience and proposals of the Organization for Economic Cooperation and Development (OECD) on the international unification of the digital tax and assess the prospects for the development of this sector of the economy, identify bottlenecks and prospects for the introduction of a similar tax in the Russian Federation.Methods: a dialectical approach to the cognition of economic regularities, which allows analyzing their development and functioning in the context of a set of objective factors, which determined the choice of the following research methods: formal-logical, economic analysis of the main trends in the construction and development of digital taxation, internal and external tax control.Results: the article discusses the current problems, foreign experience and prospects for the transition of states to the introduction of a new indirect tax – a tax on digital services. Forecasts of growth of tax revenues from digital business have been confirmed. The authors identify two trends in the international practice of digital tax development. The main methodological problems of the new tax are outlined, including the preservation of the neutrality principle, the single tax burden, the identification of the users’ location, the determination of income sources. The risks of a significant increase in the tax burden for companies and users, complication of transaction reporting systems, etc. are identified. The allocation of a tax on digital services in the Russian Federation as an independent tax is justified. The positive effect for the state and the Russian market of digital services is indicated.Scientific novelty: the paper summarizes the foreign experience of electronic economy taxation, for the first time justifies the need to adapt the national version of the digital tax, taking into account the peculiarity of the modern political and economic situation; despite the Russian support of the OECD directive, the conclusion is made about the need to segment support for the web economy with an alternative in the form of penalties for abuse in this sphere. This approach can be used as a potential opportunity to protect the national digital services market from international competition and pressure.Practical significance: the main provisions and conclusions of the article can be used in scientific, pedagogical and law enforcement activities when considering issues related to the management of the digital economy, ensuring fairness and neutrality of taxation.

Anthony O. Garuba, Dominic Eichie

Henny Rakhmawati, Mochamad Alvin Hendrawanto, Kepatuhan Wajib et al.

AbstractThe objectives of this study are to identify and explain the Regional Revenue Agency'sefforts to address the issue of boarding taxes, as well as the taxpayers' internal andexternal obstacles to paying the boarding tax.The qualitative descriptive researchmethod is utilized in this study.Five boarding houses serve as the sample for theresearch, which takes place in the Tulungagung Regency's Regional Revenue Agency(BAPENDA).This study's data sources, which include both primary and secondarydata.There are three methods for gathering data: observation, interviews, anddocumentation.Data reduction, data presentation, and conclusions were the three stagesof data analysis.According to the findings of this study, taxpayer compliance withpaying the boarding house tax in Tulungagung District can be broadly categorized asrising annually.The taxpayers' lack of awareness and lack of socialization from the lessevenly distributed BAPENDA are both internal and external constraints.Throughdoor-to-door data collection and socialization, BAPENDA hopes to collect taxpayer dataonline through WEB, which makes taxation more understandable for potentialtaxpayers. AbstrakPenelitian ini bertujuan untuk mengetahui dan menjelaskan kepatuhan wajibpajak kos di Kecamatan Tulunaggung penelitian ini menggunakan pendekatanpenelitian deskriptif kualitatif. Lokasi penelitian di Badan Pendapatan Daerah(BAPENDA) Kabupaten Tulungagung dengan sampel 5 kos-kosan. Sumberdata dalam penelitian ini menggunakan data primer dan sekunder.Pengumpulan data dilakukan dengan 3 cara, yaitu observasi, wawancara, dandokumentasi. Analisis data dilakukan dengan beberapa tahap, yaitu reduksidata, penyajian data, dan kesimpulan. Hasil penelitian ini adalah kepatuhanwajib pajak dalam membayar pajak kos- kosan di Kecamatan Tulungagungsudah bisa di kategorikan cukup baik dengan adanya peningkatan dari tahunke tahun. Kendala internal dan eksternal wajib pajak yaitu kurangnyakesadaran wajib pajak, kurangnya sosialisasi dari pihak BAPENDA yangkurang merata. Upaya dari BAPENDA yaitu melakukan pendataan denganlangkah door to door, dan melakukan sosialisasi dengan tujuan pendataanwajib pajak dengan berbasis online, seperti WEB yang memudahkan calonwajib pajak memahami tentang perpajakan.

Taejin Kim, Shubhranshu Singh, Nikhil Madaan et al.

Federated learning allows for clients in a distributed system to jointly train a machine learning model. However, clients' models are vulnerable to attacks during the training and testing phases. In this paper, we address the issue of adversarial clients performing "internal evasion attacks": crafting evasion attacks at test time to deceive other clients. For example, adversaries may aim to deceive spam filters and recommendation systems trained with federated learning for monetary gain. The adversarial clients have extensive information about the victim model in a federated learning setting, as weight information is shared amongst clients. We are the first to characterize the transferability of such internal evasion attacks for different learning methods and analyze the trade-off between model accuracy and robustness depending on the degree of similarities in client data. We show that adversarial training defenses in the federated learning setting only display limited improvements against internal attacks. However, combining adversarial training with personalized federated learning frameworks increases relative internal attack robustness by 60% compared to federated adversarial training and performs well under limited system resources.

W. Warsito, Palupi Lindiasari Samputra

Corona Virus Disease 2019 (Covid-19) pandemic has had a major negative impact on economic and taxation aspects. Indonesia’s economic growth was only 2.97 percent (first quarter of 2020), then became -5.32 percent (second quarter of 2020). Economic downturn increases the potential loss of tax revenue as the main source of state revenue so that an appropriate strategy and policies are needed to anticipate it. The objectives of this study are (1) projecting potential loss of Corporate Income Tax (CIT), Employee Income Tax (EIT), and Domestic Value Added Tax (VAT) in 2020, and (2) formulating tax policy strategy to anticipate the impact of the Covid-19 pandemic from the national resilience perspective. This study uses Seasonal Autoregressive Integrated Moving Average method to project the potential loss of CIT, EIT, and Domestic VAT in 2020 and Strengths, Weakness, Opportunities, and Threats (SWOT) analysis to formulate tax policy strategies. The results showed a potential loss of CIT, EIT, and Domestic VAT income in 2020 amounted to IDR71.748.166.578.327 (10,41 percent). This potential loss is due to decreased consumption, economic downturn, and tax incentive policies. The government can implement several tax policy strategies to anticipate the impact of the Covid-19 pandemic, namely tax priority strategies, internal strengthening strategies, collaboration and supervision strategies, and support and economic recovery strategies. Keywords: SARIMA, SWOT, tax policy strategy, the Covid-19 pandemic, national resilience Abstrak Pandemi Corona Virus Disease 2019 (Covid-19) menimbulkan dampak negatif yang besar terhadap aspek ekonomi dan perpajakan. Pertumbuhan ekonomi Indonesia pada kuartal I 2020 hanya sebesar 2,97 persen kemudian menjadi -5,32 persen pada kuartal II 2020. Perlambatan ekonomi meningkatkan potensi kehilangan penerimaan pajak sebagai sumber utama pendapatan negara sehingga diperlukan strategi dan kebijakan yang tepat untuk mengantisipasinya. Tujuan penelitian ini adalah (1) memproyeksikan potensi kehilangan penerimaan Pajak Penghasilan (PPh) Badan, PPh Pasal 21, dan Pajak Pertambahan Nilai Dalam Negeri (PPN DN) tahun 2020, dan (2) merumuskan strategi kebijakan pajak untuk mengantisipasi dampak pandemi Covid-19 dengan menggunakan perspektif ketahanan nasional. Penelitian ini menggunakan metode Seasonal Autoregressive Integrated Moving Average untuk memproyeksikan potensi kehilangan penerimaan PPh Badan, PPh Pasal 21, dan PPN DN tahun 2020 dan analisis Strengths, Weakness, Opportunities, and Threat s (SWOT) untuk merumuskan strategi kebijakan pajak. Hasil penelitian menunjukkan potensi kehilangan penerimaan PPh Badan, PPh Pasal 21, dan PPN DN tahun 2020 adalah sebesar Rp71.748.166.578.327 (10,41 persen). Potensi kehilangan penerimaan pajak tersebut disebabkan oleh penurunan konsumsi, perlambatan ekonomi, dan kebijakan insentif pajak. Pemerintah dapat menjalankan beberapa alternatif strategi kebijakan pajak dalam rangka mengantisipasi dampak pandemi Covid-19, yaitu strategi prioritas pajak, strategi penguatan internal, strategi kolaborasi dan pengawasan, serta strategi dukungan dan pemulihan ekonomi. Kata kunci: SARIMA, SWOT, strategi kebijakan pajak, pandemi Covid-19, ketahanan nasional

Raphael Bensadoun, Shir Gur, Tomer Galanti et al.

Internal learning for single-image generation is a framework, where a generator is trained to produce novel images based on a single image. Since these models are trained on a single image, they are limited in their scale and application. To overcome these issues, we propose a meta-learning approach that enables training over a collection of images, in order to model the internal statistics of the sample image more effectively. In the presented meta-learning approach, a single-image GAN model is generated given an input image, via a convolutional feedforward hypernetwork $f$. This network is trained over a dataset of images, allowing for feature sharing among different models, and for interpolation in the space of generative models. The generated single-image model contains a hierarchy of multiple generators and discriminators. It is therefore required to train the meta-learner in an adversarial manner, which requires careful design choices that we justify by a theoretical analysis. Our results show that the models obtained are as suitable as single-image GANs for many common image applications, significantly reduce the training time per image without loss in performance, and introduce novel capabilities, such as interpolation and feedforward modeling of novel images.

Santiago Balseiro, Omar Besbes, Francisco Castro

A fundamental assumption in classical mechanism design is that buyers are perfect optimizers. However, in practice, buyers may be limited by their computational capabilities or a lack of information, and may not be able to perfectly optimize. This has motivated the introduction of approximate incentive compatibility (IC) as an appealing solution concept for practical mechanism design. While most of the literature focuses on the analysis of particular approximate IC mechanisms, this paper is the first to study the design of optimal mechanisms in the space of approximate IC mechanisms and to explore how much revenue can be garnered by moving from exact to approximate incentive constraints. We study the problem of a seller facing one buyer with private values and analyze optimal selling mechanisms under $\varepsilon$-incentive compatibility. We establish that the gains that can be garnered depend on the local curvature of the seller's revenue function around the optimal posted price when the buyer is a perfect optimizer. If the revenue function behaves locally like an $α$-power for $α\in (1,\infty)$, then no mechanism can garner gains higher than order $\varepsilon^{α/(2α-1)}$. This improves upon state-of-the-art results which imply maximum gains of $\varepsilon^{1/2}$ by providing the first parametric bounds that capture the impact of revenue function's curvature on revenue gains. Furthermore, we establish that an optimal mechanism needs to randomize as soon as $\varepsilon>0$ and construct a randomized mechanism that is guaranteed to achieve order $\varepsilon^{α/(2α-1)}$ additional revenues, leading to a tight characterization of the revenue implications of approximate IC constraints. Our work brings forward the need to optimize not only over allocations and payments but also over best responses, and we develop a new framework to address this challenge.

R. Carlitz, R. McLellan

Data availability has long been a challenge for scholars of authoritarian politics. However, the promotion of open government data—through voluntary initiatives such as the Open Government Partnership and soft conditionalities tied to foreign aid—has motivated many of the world’s more closed regimes to produce and publish fine-grained data on public goods provision, taxation, and more. While this has been a boon to scholars of autocracies, we argue that the politics of data production and dissemination in these countries create new challenges. Systematically missing or biased data may jeopardize research integrity and lead to false inferences. We provide evidence of such risks from Tanzania. The example also shows how data manipulation fits into the broader set of strategies that authoritarian leaders use to legitimate and prolong their rule. Comparing data released to the public on local tax revenues with verified internal figures, we find that the public data appear to significantly underestimate opposition performance. This can bias studies on local government capacity and risk parroting the party line in data form. We conclude by providing a framework that researchers can use to anticipate and detect manipulation in newly available data.

C. Olaoye, Stephen Ayodeji Ogunleye, F. Solanke

Purpose The purpose of this paper is to examine the impact of the tax audit on tax productivity in Lagos state, Nigeria. Specifically, the study analyzed trends of tax audit and tax productivity, and the impact of Desk audit, Field audit and Back-duty audit on tax productivity in Lagos state. Design/methodology/approach The study made use of both primary and secondary data. Primary data used in the study were collected with the use of questionnaires administered to 350 randomly selected staffs of Lagos state Internal Revenue Services, while secondary data used in the study were sourced from Federal Inland Revenue Service and Lagos Internal Revenue Service audit division in Lagos state over the period spanning from 2000 to 2015. Data collated in the study were analyzed descriptively using inferential methods such as unit root test, and estimation techniques such as Fully Modified Least Square (FMOLS) co-integration regression and Logit regression analysis. Findings The study revealed that Field tax audit, desk tax audit and Back duty tax audit exert a significant positive impact on tax productivity with reported estimate of 0.530454 (p=0.0044<0.05) for FIDAUD, 0.774450 (p=0.0085< 0.05) for DEKAUD, 1.244317 (p=0.0001<0.05) for BAKAUD. Research limitations/implications Relevant tax authority (RTA), tax auditors and FIRS staff members should have full knowledge of modern audit tools like Computer Aided Audit Tools (CAATs) to enhance performance and maximum tax revenue generation. Practical implications The study concluded that tax audit enhances the level of productivity of tax administration in Lagos state and that any form of tax audit has the tendency of influencing revenue accruing to the government from taxation positively. Hence, tax audit should be carried out on a routine basis to ensure that actual revenue collected is what the RTA remits to the government. Tax audit department should be given autonomy to carry out their responsibilities effectively. Social implications Tax audit should be carried out on a routine basis to ensure that actual revenue collected is what the RTA remits to the government. Tax audit department should be given autonomy to carry out their responsibilities effectively. Originality/value This tax audit and tax productivity in Lagos state, Nigeria, fulfills an identified need to study how brand-supportive behavior can be enabled.

Célia Borlido, Mai Gehrke

Boolean spaces with internal semigroups generalize profinite semigroups and are pertinent for the recognition of not-necessarily regular languages. Via recognition, the study of existential quantification in logic on words amounts to the study of certain spans of Boolean spaces with internal semigroups. In turn, these can be understood as the superposition of a span of Boolean spaces and a span of semigroups. In this note, we first study these separately. More precisely, we identify the conditions under which each of these spans gives rise to a morphism into the respective power or Vietoris construction of the corresponding structure. Combining these characterizations, we obtain such a characterization for spans of Boolean spaces with internal semigroups which we use to describe the topo-algebraic counterpart of monadic second-order existential quantification. This is closely related to a part of the earlier work on existential quantification in first-order logic on words by Gehrke, Petri\c san and Reggio. The observation that certain morphisms lift contravariantly to the appropriate power structures makes our analysis very simple.

Zachary Sunberg, Christopher Ho, Mykel Kochenderfer

Safe interaction with human drivers is one of the primary challenges for autonomous vehicles. In order to plan driving maneuvers effectively, the vehicle's control system must infer and predict how humans will behave based on their latent internal state (e.g., intentions and aggressiveness). This research uses a simple model for human behavior with unknown parameters that make up the internal states of the traffic participants and presents a method for quantifying the value of estimating these states and planning with their uncertainty explicitly modeled. An upper performance bound is established by an omniscient Monte Carlo Tree Search (MCTS) planner that has perfect knowledge of the internal states. A baseline lower bound is established by planning with MCTS assuming that all drivers have the same internal state. MCTS variants are then used to solve a partially observable Markov decision process (POMDP) that models the internal state uncertainty to determine whether inferring the internal state offers an advantage over the baseline. Applying this method to a freeway lane changing scenario reveals that there is a significant performance gap between the upper bound and baseline. POMDP planning techniques come close to closing this gap, especially when important hidden model parameters are correlated with measurable parameters.

J. Varela, L. Garcia, S. Ohdachi et al.

LHD inward-shifted configurations are unstable to resistive MHD pressure-gradient-driven modes. These modes drive sawtooth like events during LHD operation. In this work, we simulate sawtooth like activity and internal disruptions in order to improve the understanding of these relaxation events and their effect over the device efficiency to confine the plasma, with the aim to improve the LHD present and future operation scenarios minimizing or avoiding the disadvantageous MHD soft and hard limits. By solving a set of reduced non-linear resistive MHD equations, we have studied the evolution of perturbations to equilibria obtained before and after a sawtooth like event in LHD. The equilibrium $β$ value is gradually increased during the simulation until it reaches the experimental value. Sawtooth like events and internal disruption events take place in the simulation for $β_{0}$ values between $1\%$ and $1.48\%$. The main driver of the sawtooth like events is the resonant and non-resonant effect of the $(n=1, m=3)$ mode. The instability is stronger for resonant events, and they only appear when $β_{0} = 1.48 \%$. Internal disruptions are mainly driven by the $(n=1, m=2)$ mode, and they extend throughout the whole plasma core. Internal disruption events do not show up when resonant sawtooth like events are triggered.

J. Slemrod

Lars-Erik Borge, Pernille Parmer, Ragnar Torvik

Ragavendran Gopalakrishnan, Eric Bax, Krishna Prasad Chitrapura et al.

In markets for online advertising, some advertisers pay only when users respond to ads. So publishers estimate ad response rates and multiply by advertiser bids to estimate expected revenue for showing ads. Since these estimates may be inaccurate, the publisher risks not selecting the ad for each ad call that would maximize revenue. The variance of revenue can be decomposed into two components -- variance due to `uncertainty' because the true response rate is unknown, and variance due to `randomness' because realized response statistics fluctuate around the true response rate. Over a sequence of many ad calls, the variance due to randomness nearly vanishes due to the law of large numbers. However, the variance due to uncertainty doesn't diminish. We introduce a technique for ad selection that augments existing estimation and explore-exploit methods. The technique uses methods from portfolio optimization to produce a distribution over ads rather than selecting the single ad that maximizes estimated expected revenue. Over a sequence of similar ad calls, ads are selected according to the distribution. This approach decreases the effects of uncertainty and increases revenue.