Regulation

To be led by Attorney-General.

Reinforcement learning (RL) on open-ended tasks compresses an LLM's rubric-based evaluation into a scalar reward, discarding rich textual feedback and conflating responses with distinct quality profiles. We propose Experiential Learning (EL), which repurposes the feedback model from an LLM-as-a-Judge into an LLM-as-a-Coach. The coach distills its assessment of each on-policy response into transferable experiential knowledge, which conditions a teacher model and is internalized by the policy thro

We propose Token-Level Off-Policy Labeling (TOPL), an off-policy training paradigm that reframes post-training as a token-level correctness prediction task. Our key intuition is that by training the model to distinguish good and bad tokens in a response, we naturally guide the model towards generating good tokens, while avoiding the pitfalls that come with directly training the model to generate off-policy tokens. Experiments on document summarization tasks show that TOPL achieves strong out-of-

Billing estimate engine update goes wrong.

The Interior Department’s Fish and Wildlife Service revised regulations implementing the Endangered Species Act. Conservationists warn the changes could put vulnerable species at greater risk.

Public school librarians in Ohio are raising alarms about book bans and funding cuts. School librarians have been navigating challenges in their work as long as they’ve been among the stacks in their local districts. Proposed legislation to filter the reading choices students can make has brought concern, and budget reductions make some worry about […]

The White House has repeatedly condemned EU rules that govern online platforms as censorship. Now they want the world to put the critique into writing.

Diffusion policies have shown strong potential for robotic imitation learning, and recent extensions incorporate additional modalities to improve manipulation performance. However, these modalities often differ not only in information content but also in sensing rates and inference latencies. Existing multimodal diffusion policies typically rely on synchronous fusion or manually designed multi-frequency architectures, which either slow down high-frequency feedback or limit extensibility to new m

Large language model (LLM) post-training is essential for improving reasoning, adaptation, and alignment. Existing methods mainly follow two paradigms: reinforcement learning (RL) and on-policy distillation (OPD). However, RL relies on coarse-grained outcome supervision, resulting in difficult credit assignment and limited capability to acquire new knowledge. OPD, meanwhile, unconditionally matches teacher logits through KL divergence, which creates a dilemma: similar teachers provide little new

Pretrained machine learning (ML) models help developers build ML-intensive software systems without training models from scratch. However, model repositories often provide incomplete machine-readable documentation about model provenance, licenses, datasets, limitations, and external references, creating transparency and governance gaps across the AI supply chain. Artificial Intelligence Bills of Materials (AIBOMs) address these gaps by documenting AI artifacts, including models, metadata, licens

New national plan is accompanied by Labor push for digital duty of care legislation Follow our Australia news live blog for latest updates Get our breaking news email , free app or daily news podcast The use of AI in automated decision-making by government departments and agencies will be subject to tough rules under a new national plan, expected to extend to consumer protections, workplace safety and privacy. As the Albanese government grapples with the rapid growth in the use of artificial int

Tech employees in Silicon Valley are increasingly finding themselves at odds with industry executives who are spending millions to push for light-touch AI regulation. The latest disagreement is playing out in a new political spending fight between OpenAI’s rank-and-file employees and the firm’s co-founder and president Greg Brockman. A group of current and former OpenAI...

AI agents increasingly act within the same clinical, political, scientific, and social systems that behavioral scientists study. Evaluating these systems requires source-level diagnosis: the same behavioral pattern may arise from an agent representational substrate or from the roles, objectives, interaction structures, and governance rules that shape its expression. This Perspective proposes a diagnostic framework for AI agent behavior: layer attribution. The foundational computational layer def

Agentic computer-use RL is reported in single runs, and those numbers mislead. Using verifier-guided repair of a 35B computer-use agent (CUA) across five oracle-graded environments, we show a repaired policy's success rate is dominated by upstream variance: a variance-components decomposition across three cells (crossed data-draw $\times$ seed grid, bootstrap CIs) finds evaluation variance negligible ($σ_{\mathrm{eval}} \approx 0$) and the training-seed effect small everywhere ($\leq 10\%$); ins

Publication date: September 2026 Source: Telecommunications Policy, Volume 50, Issue 8 Author(s): José Wilmar Quintero-Peña

Publication date: September 2026 Source: Telecommunications Policy, Volume 50, Issue 8 Author(s): Menglan Luo, Yong Jiang, Yi-Shuai Ren

AI governance instruments are proliferating, and so are their difficulties. Across major jurisdictions and international bodies, reform efforts built on substantially different premises encounter a recognizably similar pattern of failure. I argue that anticipatory regulatory governance rests on three operational premises —categorical stability, epistemic accessibility, and manageable pace—and that AI’s emergence, opacity, and velocity violate all three in compound. These premises form a distinct

The Credit Assignment Problem (CAP) is fundamental to developing efficient and explainable Reinforcement Learning (RL) agents. Existing frameworks, whether relying on temporal contiguity or hindsight-conditioned reward reweighting, frequently fail to attribute properly between an agent's policy (skill) and environmental stochasticity (luck). A principled approach to CAP must isolate the true causal drivers of observed outcomes from spurious correlations and environmental randomness. We introduce

This course examines refugees, migrants, stateless people, and internally displaced persons in contemporary crisis settings. It explores how conflict, political violence, disasters, climate stress, ...

Large language model (LLM) post-training is essential for improving reasoning, adaptation, and alignment. Existing methods mainly follow two paradigms: reinforcement learning (RL) and on-policy distillation (OPD). However, RL relies on coarse-grained outcome supervision, resulting in difficult credit assignment and limited capability to acquire new knowledge. OPD, meanwhile, unconditionally matches teacher logits through KL divergence, which creates a dilemma: similar teachers provide little new

Diffusion large language models (dLLMs) are a promising alternative to autoregressive generation. However, reasoning-oriented post-training for dLLMs remains challenging. Supervised fine-tuning (SFT) for dLLMs requires dense but often off-policy masked states, while reinforcement learning (RL) relies on sparse rewards or value modeling. This paper proposes \textbf{trace-based on-policy distillation (TOPD)}, a teacher-supervised framework that transfers reasoning ability to a target dLLM without

Greater global cooperation and development-focused AI regulations are urgently needed to prevent a widening digital divide, scholars and international figures said at China’s premier AI conference on Saturday. The message aligned with a broader pitch on display at the annual World AI Conference (WAIC) in Shanghai, where Beijing is courting the Global South with an artificial intelligence package centred on training, infrastructure and a bid for greater influence over global governance. “We need.

Kevin O’Leary is making the golf course argument again. The “Shark Tank” investor, whose 40,000-acre Stratos data centre project in Utah sparked protests and a governor’s executive order, told Business Insider that AI data centres consume far less water than America’s golf courses. The comparison is technically correct today. The Golf Course Superintendents Association of […] This story continues at The Next Web

Physicists have longed for a theory of everything, a single framework that would explain every force in the universe. After the better part of a century, they’re still looking. But social science, improbably, may have beaten them to it. In 2021, three British writers — John Myers, Sam Bowman, and Ben Southwood — published an […]

Half a year since Australia restricted social media companies from providing accounts to kids under 16, signs of success are beginning to show.

At the World AI Conference in Shanghai, Xi Jinping announced 5,000 AI training slots for Global South countries and the launch of the "World Artificial Intelligence Cooperation Organization." Cooperation centers with ASEAN, the African Union, BRICS, and other alliances are planned to follow. China is systematically building a parallel AI governance structure outside Western influence. The article China's new World Artificial Intelligence Cooperation Organization is President Xi's clearest play y

Von "Ich bin st�ndig erreichbar" bis zur strikten Kein-Diensthandy-nach-Feierabend-Policy ist bei euren Erz�hlungen alles dabei. ( Arbeit , Smartphone )

It used to be *the* way to get yourself a Biglaw summer job. Not anymore. The post The Death Watch For Biglaw On-Campus Interviews Has Begun appeared first on Above the Law .

The House Ways and Means Committee unanimously advanced a bill to reform Medicare Advantage prior authorization requirements. The post Healthcare Groups Praise Unanimous Committee Approval Of MA Prior Auth Bill appeared first on Above the Law .

Understanding and optimizing for voice search is increasingly vital to online search visibility, especially for capturing the all-important local search activity. The post Voice Search For Lawyers: Why Your Firm May Already Be Losing Clients appeared first on Above the Law .

“In Central Command, we’re seeing disproportionately high traumatic brain injury rates in air defense,” said Col. Jessica Peck, a command surgeon. “Even a mild traumatic brain injury can significantly degrade effectiveness, impacting coordination, emotional regulation, spatial awareness, and decision-making.” The post Military services leaning into handheld blood testing devices to diagnose TBIs appeared first on DefenseScoop .

I remember... because I'm more than 30 days old. The post Remember When Alan Dershowitz Begged To Testify About Epstein? Well, Now He’s Not Showing Up appeared first on Above the Law .

Lots of articles about this . As usual, you can also use this squid post to talk about the security stories in the news that I haven’t covered. Blog moderation policy.

Bottom line? The ostrich law firms that keep their heads in the sand are going to lose business. The post The BigHand Legal Report: More Trouble For Ostriches appeared first on Above the Law .

So, I guess those Epshteyn documents really are relevant... The post The DOJ’s Biglaw Subpoena Explanation Raises More Questions Than It Answers appeared first on Above the Law .

Entropy control has become an effective tool in reinforcement learning (RL) of large language models (LLMs), helping balance exploration-exploitation trade-off during alignment process. Such RL paradigm is often conducted on mixtures of heterogeneous tasks, which induce distinct entropy regimes under the same policy, making global or token-level entropy regulation insufficient to corresponding heterogeneous needs of exploration. This heterogeneity further makes GRPO-style normalized advantages i

On-policy knowledge distillation transfers reasoning from large teachers to compact students, but existing approaches suffer three compounding failure modes: (i) cold-start collapse, where a fresh student assigns near-zero mass to teacher-preferred tokens; (ii) state-agnostic divergence scheduling, where time-only forward/reverse-KL interpolation ignores the student's coverage state; and (iii) binary reward sparsity, where pass/fail signals discard information from partially correct traces. We p

Ben Ginsberg says years later, the proof Trump keeps promising still hasn't materialized. The post Veteran GOP Election Lawyer Still Has One Problem With Trump’s 2020 Claims: The Evidence appeared first on Above the Law .