March 2026

Reinforcement learning significantly enhances LLM capabilities but suffers from a critical issue: length inflation, where models adopt verbosity or inefficient reasoning to maximize rewards. Prior approaches struggle to address this challenge in a general and lossless manner, primarily because additive penalties introduce a compensatory effect that creates optimization shortcuts, while heuristic gating strategies lack generality beyond binary feedback. To bridge this gap, we present Group Relative Reward Rescaling (GR$^3$), which reframes length control as a multiplicative rescaling paradigm, […]

Ver mais

Unmanned aerial vehicles (UAVs) are increasingly used to support time-critical medical supply delivery, providing rapid and flexible logistics during emergencies and resource shortages. However, effective deployment of UAV fleets requires coordination mechanisms capable of prioritizing medical requests, allocating limited aerial resources, and adapting delivery schedules under uncertain operational conditions. This paper presents a multi-agent reinforcement learning (MARL) framework for coordinating UAV fleets in stochastic medical delivery scenarios where requests vary in urgency, location, and delivery deadlines. The problem […]

Ver mais

Did the machine learning illusion go to your head? You spin up a notebook, clean a fixed set of data, and train a model until the accuracy shines. Confidence grows. The prototype is perfect, and it has excited the stakeholders. “Can this go live?” is the most dangerous question in data science. This is where most promising initiatives are cut short prematurely, or even cancelled outright, in many machine learning programs. The development of a single notebook prototype […]

Ver mais

It happened once that it was tantamount to spying into a system of an unsuspecting neighborhood. You would visit a few of the servers, check CPU or memory, and decide that nothing was wrong at the end of the day. Those days are long gone. The new environment of a cloud-native system is far more dynamic and unpredictable. Several seconds are spent before containers move in and out. Inter-cloud transversal microservices clusters. The virtual networks have a journey […]

Ver mais

Excess inventory is one of the most persistent and costly inefficiencies in modern supply chains. For many organizations, it quietly erodes working capital, inflates warehousing costs, and creates a false sense of security around product availability. The traditional approach to managing inventory, which relied on intuition, spreadsheets, and static reorder points, simply cannot keep pace with the complexity of today’s demand signals. What has changed the game significantly is the application of data-driven optimization frameworks that bring together […]

Ver mais

Higher education institutions have historically relied on retrospective reporting to understand enrollment trends and student outcomes. Admissions offices scrutinized last year’s yield rates; registrars tracked semester-to-semester retention; academic affairs teams compiled graduation statistics after the fact. While these practices produced useful summaries, they offered no predictive leverage. By the time a pattern became visible in the static data, the window for meaningful intervention had almost always slammed shut. The current pivot toward predictive modeling isn’t just a technical […]

Ver mais

Modern financial institutions process millions of trades every day, and the underlying infrastructure that supports this volume must be fast, fault-tolerant, and accurate down to the microsecond. Over the past several years, event-driven architecture has emerged as the dominant pattern for building these systems, and Apache Kafka has become the backbone of many mission-critical trade processing pipelines. This post walks through the core architectural decisions, technical patterns, and operational lessons I have gathered working on high-throughput trade systems […]

Ver mais

:::info Authors: Robert Avram Jeffrey E. Olgin Zeeshan Ahmed Louis Verreault-Julien Alvin Wan Joshua Barrios Sean Abreau Derek Wan Joseph E. Gonzalez Jean-Claude Tardif Derek Y. So Krishan Soni Geoffrey H. Tison ::: Abstract Coronary angiography is the primary procedure for diagnosis and management decisions in coronary artery disease (CAD), but ad-hoc visual assessment of angiograms has high variability. Here we report a fully automated approach to interpret angiographic coronary artery stenosis from standard coronary angiograms. Using 13,843 […]

Ver mais

There is a widening chasm in enterprise technology that is quietly undermining billions in AI investments. On one side, AI capabilities are advancing faster than most organizations can track. On the other side, most enterprises are still struggling to show meaningful returns from their AI investments. Between these two realities sits what I call the execution gap – the structural disconnect between AI capability demonstrated at the model layer and measurable enterprise transformation outcomes realized at the workflow […]

Ver mais