arXiv subjects

Stefan Schmid

Publications and source records attributed to Stefan Schmid.

At least 55 records · Page 3Linked to original sources

arXiv2025

GermanPartiesQA: Benchmarking Commercial Large Language Models and AI Companions for Political Alignment and Sycophancy

Large language models (LLMs) are increasingly shaping citizens' information ecosystems. Products incorporating LLMs, such as chatbots and AI Companions, are now widely used for decision support and information retrieval, including in sensitive domains, raising concerns about hidden biases and growing potential to shape individual decisions and public opinion. This paper introduces GermanPartiesQA, a benchmark of 418 political statements from German Voting Advice Applications across 11 elections to evaluate six commercial LLMs. We evaluate their political alignment based on role-playing experiments with political personas. Our evaluation reveals three specific findings: (1) Factual limitations: LLMs show limited ability to accurately generate factual party positions, particularly for centrist parties. (2) Model-specific ideological alignment: We identify consistent alignment patterns and the degree of political steerability for each model across temperature settings and experiments. (3) Claim of sycophancy: While models adjust to political personas during role-play, we find this reflects persona-based steerability rather than the increasingly popular, yet contested concept of sycophancy. Our study contributes to evaluating the political alignment of closed-source LLMs that are increasingly embedded in electoral decision support tools and AI Companion chatbots.

cs.CY↗

arXiv2025

Whose Personae? Synthetic Persona Experiments in LLM Research and Pathways to Transparency

Synthetic personae experiments have become a prominent method in Large Language Model alignment research, yet the representativeness and ecological validity of these personae vary considerably between studies. Through a review of 63 peer-reviewed studies published between 2023 and 2025 in leading NLP and AI venues, we reveal a critical gap: task and population of interest are often underspecified in persona-based experiments, despite personalization being fundamentally dependent on these criteria. Our analysis shows substantial differences in user representation, with most studies focusing on limited sociodemographic attributes and only 35% discussing the representativeness of their LLM personae. Based on our findings, we introduce a persona transparency checklist that emphasizes representative sampling, explicit grounding in empirical data, and enhanced ecological validity. Our work provides both a comprehensive assessment of current practices and practical guidelines to improve the rigor and ecological validity of persona-based evaluations in language model alignment research.

cs.CY↗

arXiv2025

Seeing and Knowing in the Wild: Open-domain Visual Entity Recognition with Large-scale Knowledge Graphs via Contrastive Learning

Open-domain visual entity recognition aims to identify and link entities depicted in images to a vast and evolving set of real-world concepts, such as those found in Wikidata. Unlike conventional classification tasks with fixed label sets, it operates under open-set conditions, where most target entities are unseen during training and exhibit long-tail distributions. This makes the task inherently challenging due to limited supervision, high visual ambiguity, and the need for semantic disambiguation. We propose a Knowledge-guided Contrastive Learning (KnowCoL) framework that combines both images and text descriptions into a shared semantic space grounded by structured information from Wikidata. By abstracting visual and textual inputs to a conceptual level, the model leverages entity descriptions, type hierarchies, and relational context to support zero-shot entity recognition. We evaluate our approach on the OVEN benchmark, a large-scale open-domain visual recognition dataset with Wikidata IDs as the label space. Our experiments show that using visual, textual, and structured knowledge greatly improves accuracy, especially for rare and unseen entities. Our smallest model improves the accuracy on unseen entities by 10.5% compared to the state-of-the-art, despite being 35 times smaller.

cs.CV↗

arXiv2025

The Harmonic Policy for Online Buffer Sharing is (2 + ln n)-Competitive: A Simple Proof

The problem of online buffer sharing is expressed as follows. A switch with $n$ output ports receives a stream of incoming packets. When an incoming packet is accepted by the switch, it is stored in a shared buffer of capacity $B$ common to all packets and awaits its transmission through its corresponding output port determined by its destination. Each output port transmits one packet per time unit. The problem is to find an algorithm for the switch to accept or reject a packet upon its arrival in order to maximize the total number of transmitted packets. Building on the work of Kesselman et al. (STOC 2001) on split buffer sharing, Kesselman and Mansour (TCS 2004) considered the problem of online buffer sharing which models most deployed internet switches. In their work, they presented the Harmonic policy and proved that it is $(2 + \ln n)$-competitive, which is the best known competitive ratio for this problem. The Harmonic policy unfortunately saw less practical relevance as it performs $n$ threshold checks per packets which is deemed costly in practice, especially on network switches processing multiple terabits of packets per second. While the Harmonic policy is elegant, the original proof is also rather complex and involves a lengthy matching routine along with multiple intermediary results. This note presents a simplified Harmonic policy, both in terms of implementation and proof. First, we show that the Harmonic policy can be implemented with a constant number of threshold checks per packet, matching the widely deployed \emph{Dynamic Threshold} policy. Second, we present a simple proof that shows the Harmonic policy is $(2 + \ln n)$-competitive. In contrast to the original proof, the current proof is direct and relies on a 3-partitioning of the packets.

cs.DS↗

arXiv2025

A Control-Theoretic Perspective on BBR/CUBIC Congestion-Control Competition

To understand the fairness properties of the BBR congestion-control algorithm (CCA), previous research has analyzed BBR behavior with a variety of models. However, previous model-based work suffers from a trade-off between accuracy and interpretability: While dynamic fluid models generate highly accurate predictions through simulation, the causes of their predictions cannot be easily understood. In contrast, steady-state models predict CCA behavior in a manner that is intuitively understandable, but often less accurate. This trade-off is especially consequential when analyzing the competition between BBR and traditional loss-based CCAs, as this competition often suffers from instability, i.e., sending-rate oscillation. Steady-state models cannot predict this instability at all, and fluid-model simulation cannot yield analytical results regarding preconditions and severity of the oscillation. To overcome this trade-off, we extend the recent dynamic fluid model of BBR by means of control theory. Based on this control-theoretic analysis, we derive quantitative conditions for BBR/CUBIC oscillation, identify network settings that are susceptible to instability, and find that these conditions are frequently satisfied by practical networks. Our analysis illuminates the fairness implications of BBR/CUBIC oscillation, namely by deriving and experimentally validating fairness bounds that reflect the extreme rate distributions during oscillation. In summary, our analysis shows that BBR/CUBIC oscillation is frequent and harms BBR fairness, but can be remedied by means of our control-theoretic framework.

cs.NI↗

arXiv2025

Optimizing Fronthaul Quantization for Flexible User Load in Cell-Free Massive MIMO

We investigate the physical layer (PHY) spectral efficiency and fronthaul network load of a scalable user-centric cell-free massive MIMO system. Each user-centric cluster processor responsible for cluster-level signal processing is located at one of multiple decentralized units (DUs). Thus, the radio units in the cluster must exchange data with the corresponding DU over the fronthaul. Because the fronthaul links have limited capacity, this data must be quantized before it is sent over the fronthaul. We consider a routed fronthaul network, where the cluster processor placement and fronthaul traffic routing are jointly optimized with a mixed-integer linear program. For different numbers of users in the network, we investigate the effect of fronthaul quantization rates, a system parameter computed based on rate-distortion theory. Our results show that with optimized quantization rates, the fronthaul load is quite stable for a wide range of user loads without significant PHY performance loss. This demonstrates that the cell-free massive MIMO PHY and fronthaul network are resilient to varying user densities.

cs.IT↗

arXiv2025

BSB: Towards Demand-Aware Peer Selection With XOR-based Routing

Peer-to-peer networks, as a key enabler of modern networked and distributed systems, rely on peer-selection algorithms to optimize their scalability and performance. Peer-selection methods have been studied extensively in various aspects, including routing mechanisms and communication overhead. However, many state-of-the-art algorithms are oblivious to application-specific data traffic. This mismatch between design and demand results in underutilized connections, which inevitably leads to longer paths and increased latency. In this work, we propose a novel demand-aware peer-selection algorithm, called Binary Search in Buckets (BSB). Our demand-aware approach adheres to a local and greedy XOR-based routing mechanism, ensuring compatibility with existing protocols and mechanisms. We evaluate our solution against two prior algorithms by conducting simulations on real-world and synthetic communication network traces. The results of our evaluations show that BSB can offer up to a 43% improvement compared to two selected algorithms from the literature.

cs.NI↗

arXiv2025

Communication Bias in Large Language Models: A Regulatory Perspective

Large language models (LLMs) are increasingly central to many applications, raising concerns about bias, fairness, and regulatory compliance. This paper reviews risks of biased outputs and their societal impact, focusing on frameworks like the EU's AI Act and the Digital Services Act. We argue that beyond constant regulation, stronger attention to competition and design governance is needed to ensure fair, trustworthy AI. This is a preprint of the Communications of the ACM article of the same title.

cs.CY↗

arXiv2025

Boosting Payment Channel Network Liquidity with Topology Optimization and Transaction Selection

Payment channel networks (PCNs) are a promising technology that alleviates blockchain scalability by shifting the transaction load from the blockchain to the PCN. Nevertheless, the network topology has to be carefully designed to maximise the transaction throughput in PCNs. Additionally, users in PCNs also have to make optimal decisions on which transactions to forward and which to reject to prolong the lifetime of their channels. In this work, we consider an input sequence of transactions over $p$ parties. Each transaction consists of a transaction size, source, and target, and can be either accepted or rejected (entailing a cost). The goal is to design a PCN topology among the $p$ cooperating parties, along with the channel capacities, and then output a decision for each transaction in the sequence to minimise the cost of creating and augmenting channels, as well as the cost of rejecting transactions. Our main contribution is an $\mathcal{O}(p)$ approximation algorithm for the problem with $p$ parties. We further show that with some assumptions on the distribution of transactions, we can reduce the approximation ratio to $\mathcal{O}(\sqrt{p})$. We complement our theoretical analysis with an empirical study of our assumptions and approach in the context of the Lightning Network.

cs.DC↗

arXiv2025

MultiADS: Defect-aware Supervision for Multi-type Anomaly Detection and Segmentation in Zero-Shot Learning

Precise optical inspection in industrial applications is crucial for minimizing scrap rates and reducing the associated costs. Besides merely detecting if a product is anomalous or not, it is crucial to know the distinct type of defect, such as a bent, cut, or scratch. The ability to recognize the "exact" defect type enables automated treatments of the anomalies in modern production lines. Current methods are limited to solely detecting whether a product is defective or not without providing any insights on the defect type, nevertheless detecting and identifying multiple defects. We propose MultiADS, a zero-shot learning approach, able to perform Multi-type Anomaly Detection and Segmentation. The architecture of MultiADS comprises CLIP and extra linear layers to align the visual- and textual representation in a joint feature space. To the best of our knowledge, our proposal, is the first approach to perform a multi-type anomaly segmentation task in zero-shot learning. Contrary to the other baselines, our approach i) generates specific anomaly masks for each distinct defect type, ii) learns to distinguish defect types, and iii) simultaneously identifies multiple defect types present in an anomalous product. Additionally, our approach outperforms zero/few-shot learning SoTA methods on image-level and pixel-level anomaly detection and segmentation tasks on five commonly used datasets: MVTec-AD, Visa, MPDD, MAD and Real-IAD.

cs.CV↗

arXiv2025

Enhancing Manufacturing Knowledge Access with LLMs and Context-aware Prompting

Knowledge graphs (KGs) have transformed data management within the manufacturing industry, offering effective means for integrating disparate data sources through shared and structured conceptual schemas. However, harnessing the power of KGs can be daunting for non-experts, as it often requires formulating complex SPARQL queries to retrieve specific information. With the advent of Large Language Models (LLMs), there is a growing potential to automatically translate natural language queries into the SPARQL format, thus bridging the gap between user-friendly interfaces and the sophisticated architecture of KGs. The challenge remains in adequately informing LLMs about the relevant context and structure of domain-specific KGs, e.g., in manufacturing, to improve the accuracy of generated queries. In this paper, we evaluate multiple strategies that use LLMs as mediators to facilitate information retrieval from KGs. We focus on the manufacturing domain, particularly on the Bosch Line Information System KG and the I40 Core Information Model. In our evaluation, we compare various approaches for feeding relevant context from the KG to the LLM and analyze their proficiency in transforming real-world questions into SPARQL queries. Our findings show that LLMs can significantly improve their performance on generating correct and complete queries when provided only the adequate context of the KG schema. Such context-aware prompting techniques help LLMs to focus on the relevant parts of the ontology and reduce the risk of hallucination. We anticipate that the proposed techniques help LLMs to democratize access to complex data repositories and empower informed decision-making in manufacturing settings.

cs.AI↗

arXiv2025

Democracy for DAOs: An Empirical Study of Decentralized Governance and Dynamic (Case Study Internet Computer SNS Ecosystem)

Decentralized autonomous organizations (DAOs) rely on governance mechanism without centralized leadership. This paper presents an empirical study of user behavior in governance for a variety of DAOs, ranging from DeFi to gaming, using the Internet Computer Protocol DAO framework called SNS (Service Nervous System). To analyse user engagement, we measure participation rates and frequency of proposals submission and voter approval rates. We evaluate decision duration times to determine DAO agility. To investigate dynamic aspects, we also measure metric shifts in time. We evaluate over 3,000 proposals submitted in a time frame of 20 months from 14 SNS DAOs. The selected DAO have been existing between 6 and 20 months and cover a wide spectrum of use cases, treasury sizes, and number of participants. We also compare our results for SNS DAOs with DAOs from other blockchain platforms. While approval rates are generally high for all DAOs studied, SNS DAOs show slightly more alignment. We observe that the SNS governance mechanisms and processes in ICP lead to higher activity, lower costs and faster decisions. Most importantly, in contrast to studies which report a decline in participation over time for other frameworks, SNS DAOs exhibit sustained or increasing engagement levels over time.

cs.NI↗

arXiv2025

Centroid Approximation for Byzantine-Tolerant Federated Learning

Federated learning allows each client to keep its data locally when training machine learning models in a distributed setting. Significant recent research established the requirements that the input must satisfy in order to guarantee convergence of the training loop. This line of work uses averaging as the aggregation rule for the training models. In particular, we are interested in whether federated learning is robust to Byzantine behavior, and observe and investigate a tradeoff between the average/centroid and the validity conditions from distributed computing. We show that the various validity conditions alone do not guarantee a good approximation of the average. Furthermore, we show that reaching good approximation does not give good results in experimental settings due to possible Byzantine outliers. Our main contribution is the first lower bound of $\min\{\frac{n-t}{t},\sqrt{d}\}$ on the centroid approximation under box validity that is often considered in the literature, where $n$ is the number of clients, $t$ the upper bound on the number of Byzantine faults, and $d$ is the dimension of the machine learning model. We complement this lower bound by an upper bound of $2\min\{n,\sqrt{d}\}$, by providing a new analysis for the case $n<d$. In addition, we present a new algorithm that achieves a $\sqrt{2d}$-approximation under convex validity, which also proves that the existing lower bound in the literature is tight. We show that all presented bounds can also be achieved in the distributed peer-to-peer setting. We complement our analytical results with empirical evaluations in federated stochastic gradient descent and federated averaging settings.

cs.LG↗

arXiv2025

Moneros Decentralized P2P Exchanges: Functionality, Adoption, and Privacy Risks

Privacy-focused cryptocurrencies like Monero remain popular, despite increasing regulatory scrutiny that has led to their delisting from major centralized exchanges. The latter also explains the recent popularity of decentralized exchanges (DEXs) with no centralized ownership structures. These platforms typically leverage peer-to-peer (P2P) networks, promising secure and anonymous asset trading. However, questions of liability remain, and the academic literature lacks comprehensive insights into the functionality, trading activity, and privacy claims of these P2P platforms. In this paper, we provide an early systematization of the current landscape of decentralized peer-to-peer exchanges within the Monero ecosystem. We examine several recently developed DEX platforms, analyzing their popularity, functionality, architectural choices, and potential weaknesses. We further identify and report on a privacy vulnerability in the recently popularized Haveno exchange, demonstrating that certain Haveno trades could be detected, allowing transactions to be linked across the Monero and Bitcoin blockchains. We hope that our findings can nourish the discussion in the research community about more secure designs, and provide insights for regulators.

cs.CR↗

arXiv2025

An Almost Tight Lower Bound for Plurality Consensus with Undecided State Dynamics in the Population Protocol Model

We revisit the majority problem in the population protocol communication model, as first studied by Angluin et al. (Distributed Computing 2008). We consider a more general version of this problem known as plurality consensus, which has already been studied intensively in the literature. In this problem, each node in a system of $n$ nodes, has initially one of $k$ different opinions, and they need to agree on the (relative) majority opinion. In particular, we consider the important and intensively studied model of Undecided State Dynamics. Our main contribution is an almost tight lower bound on the stabilization time: we prove that there exists an initial configuration, even with bias $Δ= ω(\sqrt{n\log n})$, where stabilization requires $Ω(kn\log \frac {\sqrt n} {k \log n})$ interactions, or equivalently, $Ω(k\log \frac {\sqrt n} {k \log n})$ parallel time for any $k = o\left(\frac {\sqrt n}{\log n}\right)$. This bound is tight for any $ k \le n^{\frac 1 2 - ε}$, where $ε>0$ can be any small constant, as Amir et al.~(PODC'23) gave a $O(k\log n)$ parallel time upper bound for $k = O\left(\frac {\sqrt n} {\log ^2 n}\right)$.

cs.DC↗

arXiv2025

Brief Announcement: Minimizing Energy Solves Relative Majority with a Cubic Number of States in Population Protocols

This paper revisits a fundamental distributed computing problem in the population protocol model. Provided $n$ agents each starting with an input color in $[k]$, the relative majority problem asks to find the predominant color. In the population protocol model, at each time step, a scheduler selects two agents that first learn each other's states and then update their states based on what they learned. We present the \textsc{Circles} protocol that solves the relative majority problem with $k^3$ states. It is always-correct under weakly fair scheduling. Not only does it improve upon the best known upper bound of $O(k^7)$, but it also shows a strikingly simpler design inspired by energy minimization in chemical settings.

cs.DC↗

arXiv2025

Vermilion: A Traffic-Aware Reconfigurable Optical Interconnect with Formal Throughput Guarantees

The increasing gap between datacenter traffic volume and the capacity of electrical switches has driven the development of reconfigurable network designs utilizing optical circuit switching. Recent advancements, particularly those featuring periodic fixed-duration reconfigurations, have achieved practical end-to-end delays of just a few microseconds. However, current designs rely on multi-hop routing to enhance utilization, which can lead to a significant reduction in worst-case throughput and added overhead from congestion control and routing complexity. These factors pose significant operational challenges for the large-scale deployment of these technologies. We present Vermilion, a reconfigurable optical interconnect that breaks the throughput barrier of existing periodic reconfigurable networks, without the need for multi-hop routing -- thus eliminating congestion control and simplifying routing to direct communication. Vermilion adopts a traffic-aware approach while retaining the simplicity of periodic fixed-duration reconfigurations, similar to RotorNet. We formally establish throughput bounds for Vermilion, demonstrating that it achieves at least $33\%$ more throughput in the worst-case compared to existing designs. The key innovation of Vermilion is its short traffic-aware periodic schedule, derived using a matrix rounding technique. This schedule is then combined with a traffic-oblivious periodic schedule to efficiently manage any residual traffic. Our evaluation results support our theoretical findings, revealing significant performance gains for datacenter workloads.

cs.NI↗

arXiv2025

Approximate Agreement Algorithms for Byzantine Collaborative Learning

In Byzantine collaborative learning, $n$ clients in a peer-to-peer network collectively learn a model without sharing their data by exchanging and aggregating stochastic gradient estimates. Byzantine clients can prevent others from collecting identical sets of gradient estimates. The aggregation step thus needs to be combined with an efficient (approximate) agreement subroutine to ensure convergence of the training process. In this work, we study the geometric median aggregation rule for Byzantine collaborative learning. We show that known approaches do not provide theoretical guarantees on convergence or gradient quality in the agreement subroutine. To satisfy these theoretical guarantees, we present a hyperbox algorithm for geometric median aggregation. We practically evaluate our algorithm in both centralized and decentralized settings under Byzantine attacks on non-i.i.d. data. We show that our geometric median-based approaches can tolerate sign-flip attacks better than known mean-based approaches from the literature.

cs.LG↗