PAPER PLAINE

Researchers compared three methods for identifying informed traders on decentralized prediction markets and found they actually measure different things — not competing versions of the same measurement. One method flags accounts with consistent winning streaks, another identifies accounts behaving suspiciously over time, and a third measures how much information leaked into individual markets before public announcement. Using all three together catches more genuine insider traders than any single method alone.

A new method called ActCam lets filmmakers generate videos where they control both how an actor moves and where the camera points—without needing to train a custom AI model. By carefully layering pose and depth information at different stages of video generation, the system maintains geometric consistency and produces results that human raters prefer, especially when the camera makes large jumps to new angles.

A new framework called CLAD trains security systems across thousands of IoT devices while keeping data private and handling the reality that most collected data comes without labels. It achieves 30% better detection of network attacks than existing methods while using half the communication bandwidth, even when 80% of the data lacks security labels.

Researchers developed SNAPO, a method that trains neural networks to make sequential decisions in complex systems while simultaneously computing how sensitive those decisions are to different inputs and conditions. Unlike existing approaches that either solve small problems slowly or train fast but blind, SNAPO trains a policy in minutes while automatically generating thousands of sensitivity measurements at essentially no extra cost — a single backward pass produces both the training signal and all the risk metrics.

A new training method called StraTA helps large language models work better as decision-making agents by having them sketch out a high-level strategy before taking action. On three real-world task environments, the approach achieved success rates above 93% on some benchmarks and needed fewer training examples than existing methods.

Researchers discovered that a common way of testing image-decomposition algorithms on the MPI Sintel dataset inflates performance scores by 1.6 to 2.0 decibels because spatially similar frames from the same scene leak into both training and test sets. Using the correct evaluation method—splitting by scene rather than by frame—reveals that past reported results were significantly overstated, and the team proposes a new model that estimates uncertainty separately for different image components, allowing it to identify and filter out unreliable pixels with 77% error reduction.

Physicists have designed a way to control magnetic behavior in one material by attaching a quantum magnetic layer next to it. The quantum layer acts like a bath that damps and drives the classical magnetic material, creating new ways to tune how magnetic waves, domain walls, and skyrmions (tiny magnetic vortices) move and disappear. This could let engineers manipulate magnetic dynamics without direct electrical or magnetic contact.

When multiple AI agents work together on a task, their individual instructions (prompts) need to work well not just in isolation, but as a coordinated system. A new framework called MASPO automatically improves these prompts by testing how well each agent's output helps the next agent succeed, rather than optimizing each agent separately. Tests across six different tasks show this approach outperforms existing methods by an average of 2.9 percentage points.

New York City's 311 complaint system can't keep up with incoming calls, causing longer waits and worse service in poorer neighborhoods. Researchers built an AI system that routes complaints more intelligently—by learning that neighborhoods with repeated complaints actually need faster action, not just those with the most calls. The system reduced unfair service gaps while handling more complaints without replacing human staff.

Large language models tend to give disproportionate attention to initial tokens—a problem called "attention sink"—because of how they aggregate information and process data through their internal layers. Researchers traced this to a specific structural imbalance: early neurons create inconsistent signal strengths that force the model to anchor attention to the first token as a stabilizing mechanism. They proved this causal chain by deliberately triggering attention sinks at different positions, then tested a simple architectural fix that balanced the signals during training and sped up model convergence.

AI systems that automate computer tasks often fail when screens are high-resolution or crowded with interface elements. A new technique called BAMI improves accuracy without requiring retraining—boosting one model's performance on a challenging benchmark from 52% to 58%—by breaking down the task into simpler steps and filtering out confusing options.

Transformers work well for predicting time series, but researchers wanted to understand how—specifically whether they use the same clever internal trick (called superposition) that makes them powerful for language. By examining a transformer trained on forecasting, they found transformers actually keep things simple: they don't compress multiple patterns into the same neurons, and they ignore most of their hidden layers when making predictions. This helps explain why straightforward linear models stay competitive with far more complex transformer models.