Proceedings of Machine Learning Research | The Proceedings of Machine Learning Research (formerly JMLR Workshop and Conference Proceedings) is a series aimed specifically at publishing machine learning research presented at workshops and conferences. Each volume is separately titled and associated with a particular workshop or conference. Volumes are published online on the PMLR web site. The Series Editors are Neil D. Lawrence and Mark Reid.
The Proceedings of Machine Learning Research (formerly JMLR Workshop and Conference Proceedings) is a series aimed specifically at publishing machine learning research presented at workshops and conferences. Each volume is separately titled and associated with a particular workshop or conference....
This academic article has relevance to Intellectual Property practice area, particularly in the context of copyright law, as it mentions that authors retain copyright for their machine learning research papers published in the Proceedings of Machine Learning Research series. The article also highlights the series' publication process and guidelines, which may be of interest to IP practitioners advising clients in the field of machine learning and artificial intelligence. Additionally, the reissue series initiative may raise interesting IP considerations regarding the republication of previously published works.
**Jurisdictional Comparison and Analytical Commentary:** The publication model of the Proceedings of Machine Learning Research (PMLR) series, which allows authors to retain copyright while publishing machine learning research papers online, has significant implications for Intellectual Property (IP) practice in various jurisdictions. In the United States, the "fair use" doctrine (17 U.S.C. § 107) may protect the online publication of these research papers, but the scope of fair use is subject to interpretation. In contrast, under Korean copyright law (Article 27 of the Copyright Act), the online publication of research papers may be considered a "public performance" or "communication to the public," which requires permission from the copyright holder. Internationally, the Berne Convention for the Protection of Literary and Artistic Works (Article 8) and the World Intellectual Property Organization (WIPO) Copyright Treaty (Article 6) provide a framework for copyright protection, but the specific implementation of these treaties varies across countries. The PMLR series' approach to author retention of copyright and online publication may be seen as aligning with the principles of open access and the "green road" to open access, which is gaining traction globally. **Implications Analysis:** The PMLR series' publication model has several implications for IP practice: 1. **Author retention of copyright**: By allowing authors to retain copyright, the PMLR series promotes author autonomy and control over their work, which may be beneficial for researchers in
As a Patent Prosecution & Infringement Expert, I analyzed the article's implications for practitioners in the field of intellectual property and patent law. The article discusses the Proceedings of Machine Learning Research (PMLR), a series that publishes machine learning research papers. This series is relevant to patent practitioners as it may be considered prior art in patent applications related to machine learning inventions. The PMLR series is a collection of published works that can be used to demonstrate the existence of prior art, which can impact the novelty and non-obviousness of a patent application. In the context of patent law, the PMLR series is akin to a collection of prior art references, which can be used to support patent validity and infringement analyses. For example, in the case of _KSR Int'l Co. v. Teleflex Inc._, 550 U.S. 398 (2007), the Supreme Court emphasized the importance of considering prior art in determining the non-obviousness of a patent. The PMLR series can be used to identify prior art that may have a bearing on the patentability of machine learning inventions. From a statutory perspective, the PMLR series is relevant to the patentability requirements of 35 U.S.C. § 101, which requires that an invention be novel and non-obvious. The PMLR series can be used to demonstrate the existence of prior art that may impact the novelty and non-obviousness of a patent application. In terms
Stand Tall for the Rule of Law - a Film
Based on the provided article, here's an analysis of its relevance to Intellectual Property (IP) practice area: The article discusses an event and a film, "Stand Tall for the Rule of Law," which focuses on reaffirming commitment to fundamental principles of international law and promoting human rights. However, there is no direct mention of Intellectual Property law. Nevertheless, the event's emphasis on international law and human rights may have implications for IP law, particularly in the context of global governance and the protection of intellectual property rights. In terms of key legal developments, research findings, and policy signals, the article does not provide any specific information. However, the event's focus on international law and human rights may signal a growing interest in global governance and the protection of human rights, which could potentially impact IP law and policy in the future.
The article’s impact on Intellectual Property practice is nuanced, as it primarily centers on international law reaffirmation rather than IP-specific provisions. Nonetheless, its symbolic alignment with international legal milestones—the 75th anniversary of the Genocide Convention and the Universal Declaration of Human Rights—reinforces the broader principle that legal integrity underpins all rights, including IP. In comparative perspective, the U.S. approach typically anchors IP protection in statutory codification (e.g., Lanham Act, Patent Act) and judicial precedent, while Korea emphasizes statutory harmonization with international treaties (e.g., TRIPS, WIPO) and administrative enforcement via KIPO. Internationally, the trend leans toward multilateral cooperation over unilateral enforcement, as evidenced by the ASIL summit’s emphasis on shared legal values. Thus, while the film does not alter IP doctrine, it subtly amplifies the cultural and institutional imperative that legal systems, including IP regimes, must be anchored in principled, collective governance.
As a Patent Prosecution & Infringement Expert, I must note that the provided article appears to be unrelated to the field of intellectual property law. However, I can provide a general analysis of the article's implications for practitioners in the context of international law and its potential connections to patent law. The article highlights the importance of upholding the rule of law, particularly in the context of international law and human rights. This is relevant to patent practitioners in that it underscores the need for a robust and impartial legal system to protect intellectual property rights. The Genocide Convention and the Universal Declaration of Human Rights, mentioned in the article, have implications for the protection of intellectual property rights in international trade and commerce. In the context of patent law, the concept of "rule of law" is closely tied to the idea of "rule of law" in patent prosecution, which emphasizes the importance of a fair and transparent process for granting and enforcing patents. This is reflected in statutory provisions such as 35 U.S.C. § 2, which states that the patent laws shall be administered in a manner that promotes the progress of science and useful arts. In terms of case law, the concept of "rule of law" is closely tied to the idea of "patent exhaustion," which holds that once a patent has been granted, the patentee's rights are exhausted, and the patented invention can be freely used and sold by others (see, e.g., Quanta Computer, Inc. v. LG Electronics
Assessing States’ Obligations under the UN Guiding Principles on Business and Human Rights Post-Brexit
Private economic actors wield unprecedented influence over the enjoyment of human rights, yet legal systems remain uneven in their regulation of corporate responsibility. Against this backdrop, this article examines a largely underexplored post-Brexit trajectory, the regulatory divergence in the implementation...
This article is relevant to Intellectual Property practice as it highlights regulatory divergence post-Brexit in corporate human rights accountability, a growing intersection between IP rights (especially in tech and pharma sectors) and human rights obligations. The comparative analysis of EU preventative regulation versus UK minimalist adjudication offers policy signals for stakeholders navigating cross-border IP disputes where human rights compliance intersects with corporate conduct. The focus on Northern Ireland as a hybrid regulatory space signals emerging legal complexities for IP practitioners managing jurisdictional overlaps in human rights-sensitive industries.
The article’s analysis of regulatory divergence post-Brexit offers a pertinent lens for Intellectual Property (IP) practitioners, particularly as IP rights intersect with corporate accountability and human rights obligations. While the EU’s preventative regulatory framework aligns with broader IP enforcement strategies that emphasize proactive compliance and systemic oversight, the UK’s minimalist adjudicative model reflects a reactive posture akin to certain IP dispute resolution mechanisms—both favoring adjudication over preemptive governance. Internationally, jurisdictions like South Korea exemplify a hybrid approach, integrating IP protection with human rights principles through statutory mandates and administrative oversight, thereby bridging EU and UK extremes. This comparative divergence underscores a broader tension between transnational governance legitimacy and localized implementation, influencing IP stakeholders navigating corporate responsibility frameworks globally.
The article implicates practitioners in IP and human rights law by highlighting the growing influence of private actors on human rights and the regulatory divergence between EU and UK post-Brexit approaches to corporate accountability. Practitioners should anticipate increased scrutiny of corporate conduct under evolving transnational governance frameworks, akin to the UNGPs, which may influence how human rights considerations intersect with IP rights, especially in cross-border disputes. Statutorily, this aligns with the UNGPs’ influence on domestic regulatory frameworks, while case law such as *UN Guiding Principles on Business and Human Rights* (interpreted through domestic courts’ evolving jurisprudence) may shape future litigation strategies involving corporate responsibility. Regulatory divergence underscores the need for practitioners to adapt strategies to jurisdictional nuances, particularly in jurisdictions like Northern Ireland where hybrid legal alignment creates unique compliance challenges.
Is the Electronic Trade Documents Act 2023 Sufficient to Promote the Uptake of Paperless Trading Systems?
In September 2023, the Electronic Trade Documents Act (ETDA) came into force in the UK. It aims to facilitate paperless trade by allowing certain trade documents in electronic form to have the same legal functionality as their paper counterparts. The...
The ETDA analysis is relevant to IP practice as it highlights a critical gap between legal recognition of electronic documents and the operational incentives needed to drive digitisation—specifically, the “membership requirement” remains a barrier even with statutory recognition. This signals a broader policy signal for IP stakeholders: legal frameworks alone (e.g., statutory electronic document recognition) are insufficient without harmonised governance standards to build trust and standardisation across commercial parties. The findings may inform IP strategies around digital trade agreements, contract design, and dispute resolution in electronic commerce.
The Electronic Trade Documents Act 2023 in the UK presents an interesting case study for analyzing the impact of enabling legislation on the digitization of trade. When compared to the US and Korean approaches, the UK's ETDA appears to share similarities with the US's Uniform Electronic Transactions Act (UETA), which aims to provide a uniform framework for electronic transactions. However, both the US and Korean approaches differ from the UK's ETDA in their emphasis on standardization and governance, as seen in Korea's Electronic Signature Act, which mandates the use of standardized electronic signatures. In the US, the UETA focuses on providing a framework for electronic transactions, but leaves the development of standards and governance to industry players. In contrast, Korea's Electronic Signature Act requires the use of standardized electronic signatures, which has contributed to the country's high adoption rate of electronic signatures. The UK's ETDA, on the other hand, requires the use of a reliable system to ensure that electronic trade documents are valid, but does not address the issue of standardization and governance. Internationally, the United Nations Commission on International Trade Law (UNCITRAL) has developed a model law on electronic signatures, which provides a framework for countries to develop their own laws on electronic signatures. However, the lack of standardization and governance in the UK's ETDA may hinder its effectiveness in promoting the digitization of trade. In conclusion, while the UK's ETDA shares similarities with the US's UETA, its lack of emphasis on
The ETDA 2023 addresses a critical juncture in the digitisation of trade by recognising electronic documents legally, yet practitioners should note that the article identifies a key limitation: the legislation does not fully resolve the "membership requirement," which remains a significant barrier to widespread adoption. This aligns with broader principles of statutory efficacy, echoing case law on the necessity of comprehensive frameworks to enforce intended legislative outcomes—e.g., analogous challenges in digital authentication under the UK’s Electronic Communications Act 2000. Statutorily, the ETDA’s reliance on a "reliable system" may insufficiently compel systemic change without a standardised governance mechanism, suggesting that regulatory incentives must complement legal recognition to drive uptake. Practitioners should anticipate the need for additional regulatory or contractual mechanisms to bridge this gap in incentivising digitisation.
Making Rights Fundamental: The 2022 Amendment to the 1998 ILO Declaration on Fundamental Principles and Rights at Work and its Radical Implications
What makes a right fundamental, and how does it achieve this status? This article critically examines these questions through a detailed analysis of the 2022 amendment to the 1998 ILO Declaration, which recognised the right to a safe and healthy...
This article has limited direct relevance to Intellectual Property (IP) practice area, but it has some tangential implications. The analysis focuses on the 2022 amendment to the 1998 ILO Declaration, recognizing the right to a safe and healthy working environment as a fifth fundamental right, which may have implications for labor rights and social justice. Key developments include: - The 2022 amendment to the 1998 ILO Declaration recognizing a new fundamental right to a safe and healthy working environment. - The emergence of a flexible 'amendment formula' that may lower the bar for future rights to be added. - The article's argument that the 2022 amendment makes accounts of fundamental rights under the 1998 Declaration as 'procedural' or 'enabling' untenable. Research findings suggest that the process of fundamentalization (gaining fundamental status) is influenced by internal and external actors and factors, including COVID-19. The article also highlights the predominance of constitutional-textual and rights-based justifications of the amendment, which were informed primarily by ILO Conventions. Policy signals from this article are limited to the labor rights and social justice context, but they may have broader implications for the concept of fundamentality and the process of fundamentalization in various areas of law, including IP.
The 2022 amendment to the ILO Declaration offers instructive parallels for Intellectual Property (IP) practitioners, particularly in its analysis of "fundamentalisation" and the criteria that elevate a right to a normative status. While IP rights are typically codified through statutory and treaty mechanisms rather than constitutional or rights-based frameworks, the article’s exploration of how external crises (e.g., COVID-19) influence normative recognition aligns with IP’s evolving recognition of rights in response to societal shifts—such as the expansion of moral rights or data privacy protections. From a jurisdictional perspective, the U.S. tends to anchor IP rights in statutory and contractual frameworks, whereas Korean IP law integrates a blend of statutory enforcement and constitutional principles, particularly in matters of privacy and consumer protection. Internationally, the ILO’s amendment signals a trend toward embedding rights through interpretive evolution—a mechanism that IP regimes may emulate in adapting to emerging challenges, such as AI-generated content or biotech innovations. The flexible “amendment formula” identified in the article could inspire analogous pathways for IP rights to evolve without requiring exhaustive legislative overhaul, fostering agility in rights recognition.
As a Patent Prosecution & Infringement Expert, I'll provide domain-specific expert analysis of the article's implications for practitioners, noting any case law, statutory, or regulatory connections. **Analysis:** While the article primarily focuses on labor rights and the International Labor Organization (ILO) Declaration, there are some indirect implications for intellectual property (IP) practitioners. The concept of "fundamentality" and the process of "fundamentalisation" may be applicable to IP rights, particularly in the context of human rights and social responsibility. In the IP domain, the notion of "fundamentality" could be linked to the concept of "public interest" or "social utility," which is often considered when evaluating the validity or enforceability of IP rights. For instance, in the context of patent law, the public interest may be a factor in determining the scope of patent protection or the applicability of exceptions and limitations. **Case Law, Statutory, or Regulatory Connections:** * The article's discussion on the ILO Declaration and the concept of fundamentality may be relevant to the interpretation of human rights and social responsibility in the context of IP law. For example, the European Court of Human Rights (ECHR) has considered the relationship between IP rights and human rights in cases such as _Centrum voor burgerrechten v. the Netherlands_ (2012). * The idea of "fundamentalisation" as a process of gaining normative status may be analogous to the concept of
Navigating the New Frontier: How AI Regulation is Reshaping the Global Technology Landscape
As of February 2026, the global technology landscape is undergoing a significant transformation driven by the increasing regulation of Artificial Intelligence (AI). Governments and regulatory bodies around the world are implementing new laws and guidelines to ensure the safe and...
**Relevance to Intellectual Property Practice Area:** The article highlights key legal developments in AI regulation, with implications for intellectual property (IP) practice. The increasing regulation of AI is expected to impact various industries and aspects of society, including IP rights. The article's analysis of the current state of AI regulation, its implications, and the future of the technology sector is relevant to IP practitioners who need to stay up-to-date on the evolving regulatory landscape. **Key Legal Developments, Research Findings, and Policy Signals:** 1. **Global AI Regulation:** Governments and regulatory bodies are implementing laws and guidelines to regulate AI development and deployment, reshaping the global technology landscape. 2. **EU's AI Regulation:** The proposed Artificial Intelligence Act will establish a framework for AI system development and deployment, categorizing them based on risk and imposing strict requirements on high-risk applications. 3. **FTC's AI Guidelines:** The Federal Trade Commission's guidelines on AI and machine learning emphasize the importance of transparency, explainability, and fairness in AI-driven decision-making processes. These developments signal a shift towards more stringent regulation of AI, which will likely impact IP rights, such as patentability, trade secrets, and copyright protection. IP practitioners should be aware of these changes to ensure they are adapting to the evolving regulatory landscape.
The AI regulation landscape reveals distinct jurisdictional contours that influence Intellectual Property (IP) practice globally. In the EU, the GDPR and AI Act establish a risk-based framework that directly intersects with IP by regulating data-derived AI outputs and imposing transparency obligations on algorithmic innovation, thereby affecting patent eligibility and trade secret protection. The U.S. approach, led by the FTC’s enforcement of consumer protection principles through transparency and fairness mandates, operates more through administrative oversight than statutory codification, creating a flexible but less predictable IP enforcement environment. Internationally, the WIPO-led discourse on AI-generated content as subject to IP rights (e.g., in the Draft WIPO Interdisciplinary Study) signals a converging trend toward recognizing AI-generated outputs as protectable under existing IP regimes, albeit with jurisdictional variance in application. These divergent models—EU’s statutory codification, U.S.’s administrative pragmatism, and WIPO’s normative convergence—create layered compliance challenges for multinational IP stakeholders, necessitating adaptive strategies across jurisdictions.
As a Patent Prosecution & Infringement Expert, the implications of AI regulation for practitioners are significant. First, the emergence of comprehensive AI frameworks, such as the EU’s Artificial Intelligence Act and GDPR, may influence patent eligibility and claim drafting, particularly for AI-related inventions that touch on data privacy or ethical considerations. Practitioners should anticipate heightened scrutiny of claims involving AI applications in regulated domains, aligning with statutory frameworks like the FTC’s focus on transparency and fairness. Second, as seen in precedents like Alice Corp. v. CLS Bank, claims that lack a tangible, non-abstract improvement risk invalidity, making it critical to anchor AI innovations in concrete technical solutions rather than purely algorithmic processes. These regulatory shifts thus necessitate a more nuanced approach to claim construction and prosecution strategy.
Final 2 days to save up to $500 on your TechCrunch Disrupt 2026 ticket
Ticket discounts of up to $500 will end tomorrow, April 10, at 11:59 p.m. PT. After that, prices for TechCrunch Disrupt 2026 go up again. Miss this, and you’ll be paying more for the same access to one of the...
Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook
arXiv:2604.06210v1 Announce Type: new Abstract: As LLMs are globally deployed, aligning their cultural value orientations is critical for safety and user engagement. However, existing benchmarks face the Construct-Composition-Context ($C^3$) challenge: relying on discriminative, multiple-choice formats that probe value knowledge rather...
When to Call an Apple Red: Humans Follow Introspective Rules, VLMs Don't
arXiv:2604.06422v1 Announce Type: new Abstract: Understanding when Vision-Language Models (VLMs) will behave unexpectedly, whether models can reliably predict their own behavior, and if models adhere to their introspective reasoning are central challenges for trustworthy deployment. To study this, we introduce...
Depression Detection at the Point of Care: Automated Analysis of Linguistic Signals from Routine Primary Care Encounters
arXiv:2604.06193v1 Announce Type: new Abstract: Depression is underdiagnosed in primary care, yet timely identification remains critical. Recorded clinical encounters, increasingly common with digital scribing technologies, present an opportunity to detect depression from naturalistic dialogue. We investigated automated depression detection from...
A Parameter-Efficient Transfer Learning Approach through Multitask Prompt Distillation and Decomposition for Clinical NLP
arXiv:2604.06650v1 Announce Type: new Abstract: Existing prompt-based fine-tuning methods typically learn task-specific prompts independently, imposing significant computing and storage overhead at scale when deploying multiple clinical natural language processing (NLP) systems. We present a multitask prompt distillation and decomposition framework...
Scoring Edit Impact in Grammatical Error Correction via Embedded Association Graphs
arXiv:2604.06573v1 Announce Type: new Abstract: A Grammatical Error Correction (GEC) system produces a sequence of edits to correct an erroneous sentence. The quality of these edits is typically evaluated against human annotations. However, a sentence may admit multiple valid corrections,...
State-of-the-Art Arabic Language Modeling with Sparse MoE Fine-Tuning and Chain-of-Thought Distillation
arXiv:2604.06421v1 Announce Type: new Abstract: This paper introduces Arabic-DeepSeek-R1, an application-driven open-source Arabic LLM that leverages a sparse MoE backbone to address the digital equity gap for under-represented languages, and establishes a new SOTA across the entire Open Arabic LLM...
Temporally Phenotyping GLP-1RA Case Reports with Large Language Models: A Textual Time Series Corpus and Risk Modeling
arXiv:2604.06197v1 Announce Type: new Abstract: Type 2 diabetes case reports describe complex clinical courses, but their timelines are often expressed in language that is difficult to reuse in longitudinal modeling. To address this gap, we developed a textual time-series corpus...
Beyond Facts: Benchmarking Distributional Reading Comprehension in Large Language Models
arXiv:2604.06201v1 Announce Type: new Abstract: While most reading comprehension benchmarks for LLMs focus on factual information that can be answered by localizing specific textual evidence, many real-world tasks require understanding distributional information, such as population-level trends and preferences expressed across...
AE-ViT: Stable Long-Horizon Parametric Partial Differential Equations Modeling
arXiv:2604.06475v1 Announce Type: new Abstract: Deep Learning Reduced Order Models (ROMs) are becoming increasingly popular as surrogate models for parametric partial differential equations (PDEs) due to their ability to handle high-dimensional data, approximate highly nonlinear mappings, and utilize GPUs. Existing...
MICA: Multivariate Infini Compressive Attention for Time Series Forecasting
arXiv:2604.06473v1 Announce Type: new Abstract: Multivariate forecasting with Transformers faces a core scalability challenge: modeling cross-channel dependencies via attention compounds attention's quadratic sequence complexity with quadratic channel scaling, making full cross-channel attention impractical for high-dimensional time series. We propose Multivariate...
SensorPersona: An LLM-Empowered System for Continual Persona Extraction from Longitudinal Mobile Sensor Streams
arXiv:2604.06204v1 Announce Type: new Abstract: Personalization is essential for Large Language Model (LLM)-based agents to adapt to users' preferences and improve response quality and task performance. However, most existing approaches infer personas from chat histories, which capture only self-disclosed information...
Limits of Difficulty Scaling: Hard Samples Yield Diminishing Returns in GRPO-Tuned SLMs
arXiv:2604.06298v1 Announce Type: new Abstract: Recent alignment work on Large Language Models (LLMs) suggests preference optimization can improve reasoning by shifting probability mass toward better solutions. We test this claim in a resource-constrained setting by applying GRPO with LoRA to...
Invisible Influences: Investigating Implicit Intersectional Biases through Persona Engineering in Large Language Models
arXiv:2604.06213v1 Announce Type: new Abstract: Large Language Models (LLMs) excel at human-like language generation but often embed and amplify implicit, intersectional biases, especially under persona-driven contexts. Existing bias audits rely on static, embedding-based tests (CEAT, I-WEAT, I-SEAT) that quantify absolute...
From Load Tests to Live Streams: Graph Embedding-Based Anomaly Detection in Microservice Architectures
arXiv:2604.06448v1 Announce Type: new Abstract: Prime Video regularly conducts load tests to simulate the viewer traffic spikes seen during live events such as Thursday Night Football as well as video-on-demand (VOD) events such as Rings of Power. While these stress...
Illocutionary Explanation Planning for Source-Faithful Explanations in Retrieval-Augmented Language Models
arXiv:2604.06211v1 Announce Type: new Abstract: Natural language explanations produced by large language models (LLMs) are often persuasive, but not necessarily scrutable: users cannot easily verify whether the claims in an explanation are supported by evidence. In XAI, this motivates a...
Does a Global Perspective Help Prune Sparse MoEs Elegantly?
arXiv:2604.06542v1 Announce Type: new Abstract: Empirical scaling laws for language models have encouraged the development of ever-larger LLMs, despite their growing computational and memory costs. Sparse Mixture-of-Experts (MoEs) offer a promising alternative by activating only a subset of experts per...
The Rhetoric of Machine Learning
arXiv:2604.06754v1 Announce Type: new Abstract: I examine the technology of machine learning from the perspective of rhetoric, which is simply the art of persuasion. Rather than being a neutral and "objective" way to build "world models" from data, machine learning...
AgentOpt v0.1 Technical Report: Client-Side Optimization for LLM-Based Agent
arXiv:2604.06296v1 Announce Type: new Abstract: AI agents are increasingly deployed in real-world applications, including systems such as Manus, OpenClaw, and coding agents. Existing research has primarily focused on \emph{server-side} efficiency, proposing methods such as caching, speculative execution, traffic scheduling, and...
TalkLoRA: Communication-Aware Mixture of Low-Rank Adaptation for Large Language Models
arXiv:2604.06291v1 Announce Type: new Abstract: Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning of Large Language Models (LLMs), and recent Mixture-of-Experts (MoE) extensions further enhance flexibility by dynamically combining multiple LoRA experts. However, existing MoE-augmented LoRA methods assume that experts operate independently,...
Asymptotic-Preserving Neural Networks for Viscoelastic Parameter Identification in Multiscale Blood Flow Modeling
arXiv:2604.06287v1 Announce Type: new Abstract: Mathematical models and numerical simulations offer a non-invasive way to explore cardiovascular phenomena, providing access to quantities that cannot be measured directly. In this study, we start with a one-dimensional multiscale blood flow model that...
FlowAdam: Implicit Regularization via Geometry-Aware Soft Momentum Injection
arXiv:2604.06652v1 Announce Type: new Abstract: Adaptive moment methods such as Adam use a diagonal, coordinate-wise preconditioner based on exponential moving averages of squared gradients. This diagonal scaling is coordinate-system dependent and can struggle with dense or rotated parameter couplings, including...
LLM-Augmented Knowledge Base Construction For Root Cause Analysis
arXiv:2604.06171v1 Announce Type: new Abstract: Communications networks now form the backbone of our digital world, with fast and reliable connectivity. However, even with appropriate redundancy and failover mechanisms, it is difficult to guarantee "five 9s" (99.999 %) reliability, requiring rapid...
The Depth Ceiling: On the Limits of Large Language Models in Discovering Latent Planning
arXiv:2604.06427v1 Announce Type: new Abstract: The viability of chain-of-thought (CoT) monitoring hinges on models being unable to reason effectively in their latent representations. Yet little is known about the limits of such latent reasoning in LLMs. We test these limits...