Multilevel Training for Kolmogorov Arnold Networks
arXiv:2603.04827v1 Announce Type: new Abstract: Algorithmic speedup of training common neural architectures is made difficult by the lack of structure guaranteed by the function compositions inherent to such networks. In contrast to multilayer perceptrons (MLPs), Kolmogorov-Arnold networks (KANs) provide more...
Missingness Bias Calibration in Feature Attribution Explanations
arXiv:2603.04831v1 Announce Type: new Abstract: Popular explanation methods often produce unreliable feature importance scores due to missingness bias, a systematic distortion that arises when models are probed with ablated, out-of-distribution inputs. Existing solutions treat this as a deep representational flaw...
Why Is RLHF Alignment Shallow? A Gradient Analysis
arXiv:2603.04851v1 Announce Type: new Abstract: Why is safety alignment in LLMs shallow? We prove that gradient-based alignment inherently concentrates on positions where harm is decided and vanishes beyond. Using a martingale decomposition of sequence-level harm, we derive an exact characterization...
Differential Privacy in Two-Layer Networks: How DP-SGD Harms Fairness and Robustness
arXiv:2603.04881v1 Announce Type: new Abstract: Differentially private learning is essential for training models on sensitive data, but empirical studies consistently show that it can degrade performance, introduce fairness issues like disparate impact, and reduce adversarial robustness. The theoretical underpinnings of...
FedAFD: Multimodal Federated Learning via Adversarial Fusion and Distillation
arXiv:2603.04890v1 Announce Type: new Abstract: Multimodal Federated Learning (MFL) enables clients with heterogeneous data modalities to collaboratively train models without sharing raw data, offering a privacy-preserving framework that leverages complementary cross-modal information. However, existing methods often overlook personalized client performance...
BandPO: Bridging Trust Regions and Ratio Clipping via Probability-Aware Bounds for LLM Reinforcement Learning
arXiv:2603.04918v1 Announce Type: new Abstract: Proximal constraints are fundamental to the stability of the Large Language Model reinforcement learning. While the canonical clipping mechanism in PPO serves as an efficient surrogate for trust regions, we identify a critical bottleneck: fixed...
Generative AI in legal education: a two-year experiment with ChatGPT
Justices poised to adopt exceptions to federal criminal defendants’ appellate waivers
The Supreme Court heard oral argument on Tuesday in Hunter v. United States about what exceptions exist to federal defendants’ waivers of their right to appeal. The justices seemed poised […]The postJustices poised to adopt exceptions to federal criminal defendants’...
Birthright citizenship: the exceptions provide the rule
The battle over birthright citizenship is a battle over its exceptions. The 14th Amendment’s first sentence proudly proclaims that “[a]ll persons born . . . in the United States, and subject to the jurisdiction […]The postBirthright citizenship: the exceptions provide...
The emergency docket’s critics have it backwards
Ratio Decidendi is a recurring series by Stephanie Barclay exploring the reasoning – from practical considerations to deep theory – behind our nation’s most consequential constitutional decisions. Last Monday, the […]The postThe emergency docket’s critics have it backwardsappeared first onSCOTUSblog.
Will the mystery of the Dobbs leak ever be solved?
Justice Clarence Thomas’ virtual appearance last week at a legal conference in Washington, D.C. brought renewed attention to court security. Thomas had originally planned to attend in person, but he […]The postWill the mystery of the Dobbs leak ever be...
SCOTUStoday for Friday, March 6
On this day in 1857, the Supreme Court released its opinion in Dred Scott v. Sandford, holding that Scott, an enslaved man who spent time in free territory, was not […]The postSCOTUStoday for Friday, March 6appeared first onSCOTUSblog.
Syrian nationals urge Supreme Court to keep ruling in place allowing them to stay in the United States
A group of Syrian nationals urged the Supreme Court on Thursday to leave in place a ruling by a federal judge in New York City that allows them to remain […]The postSyrian nationals urge Supreme Court to keep ruling in...
Court grapples with whether federal law supersedes negligent hiring claims against freight brokers
Updated on March 6 at 10:50 a.m. The Supreme Court on Wednesday heard argument in Montgomery v. Caribe Transport II, LLC, a case on whether federal law prevents state law […]The postCourt grapples with whether federal law supersedes negligent hiring...
Supreme Court rules that New Jersey Transit can be sued in other states
The Supreme Court on Wednesday ruled in Galette v. New Jersey Transit Corporation that two men who were seriously injured in New York and Pennsylvania by buses operated by New […]The postSupreme Court rules that New Jersey Transit can be...
AI Now Institute
AI Now Institute | 19,196 followers on LinkedIn. The AI Now Institute produces diagnosis and actionable policy research on artificial intelligence.
Musk fails to block California data disclosure law he fears will ruin xAI
Musk can't convince judge public doesn’t care about where AI training data comes from.
Tech industry is in tariff hell, even if refunds are automated
Trade groups urge court to create a simple blueprint for tariff refunds.
Trump gets data center companies to pledge to pay for power generation
With no enforcement and questionable economics, it may not make a difference.
Anthropic’s Pentagon deal is a cautionary tale for startups chasing federal contracts
The Pentagon has officially designated Anthropic a supply-chain risk after the two failed to agree on how much control the military should have over its AI models, including its use in autonomous weapons and mass domestic surveillance. As Anthropic’s $200...
Anthropic vs. the Pentagon, the SaaSpocalypse, and why competitions is good, actually
The Pentagon has officially designated Anthropic a supply-chain risk after the two failed to agree on how much control the military should have over its AI models, including its use in autonomous weapons and mass domestic surveillance. As Anthropic’s $200...
DiligenceSquared uses AI, voice agents to make M&A research affordable
Instead of relying on expensive management consultants, the startup uses AI voice agents to conduct interviews with customers of the companies the PE firms are considering buying.
Anthropic CEO Dario Amodei could still be trying to make a deal with Pentagon
Anthropic's $200 million contract with the Department of Defense broke down due to disagreements over giving the military unrestricted access to its AI.
A Dual-Helix Governance Approach Towards Reliable Agentic AI for WebGIS Development
arXiv:2603.04390v1 Announce Type: new Abstract: WebGIS development requires rigor, yet agentic AI frequently fails due to five large language model (LLM) limitations: context constraints, cross-session forgetting, stochasticity, instruction failure, and adaptation rigidity. We propose a dual-helix governance framework reframing these...
One Bias After Another: Mechanistic Reward Shaping and Persistent Biases in Language Reward Models
arXiv:2603.03291v1 Announce Type: cross Abstract: Reward Models (RMs) are crucial for online alignment of language models (LMs) with human preferences. However, RM-based preference-tuning is vulnerable to reward hacking, whereby LM policies learn undesirable behaviors from flawed RMs. By systematically measuring...
Language Model Goal Selection Differs from Humans' in an Open-Ended Task
arXiv:2603.03295v1 Announce Type: cross Abstract: As large language models (LLMs) get integrated into human decision-making, they are increasingly choosing goals autonomously rather than only completing human-defined ones, assuming they will reflect human preferences. However, human-LLM similarity in goal selection remains...
TTSR: Test-Time Self-Reflection for Continual Reasoning Improvement
arXiv:2603.03297v1 Announce Type: cross Abstract: Test-time Training enables model adaptation using only test questions and offers a promising paradigm for improving the reasoning ability of large language models (LLMs). However, it faces two major challenges: test questions are often highly...
TATRA: Training-Free Instance-Adaptive Prompting Through Rephrasing and Aggregation
arXiv:2603.03298v1 Announce Type: cross Abstract: Large Language Models (LLMs) have improved substantially alignment, yet their behavior remains highly sensitive to prompt phrasing. This brittleness has motivated automated prompt engineering, but most existing methods (i) require a task-specific training set, (ii)...
Developing an AI Assistant for Knowledge Management and Workforce Training in State DOTs
arXiv:2603.03302v1 Announce Type: cross Abstract: Effective knowledge management is critical for preserving institutional expertise and improving the efficiency of workforce training in state transportation agencies. Traditional approaches, such as static documentation, classroom-based instruction, and informal mentorship, often lead to fragmented...