Papers

Listen to this article.

Problem

Current large language models (LLMs) often excel at isolated tasks like next-token prediction, but struggle to truly understand and interact with the world in a unified way. This paper addresses the need for more holistic AI systems that can reason about states, predict transitions, and ultimately act upon the world in a coherent manner.

Method

The authors introduce “Orca,” a world foundation model designed to learn a single, unified representation of the world – a “world latent space.” This is achieved through a novel approach called Next-State-Prediction modeling, moving away from traditional next-token prediction towards forecasting how states evolve over time. Crucially, Orca employs two learning paradigms:

Listen to this article.

Problem

Real-time video editing, especially in interactive and augmented reality (AR) scenarios, faces significant challenges. Existing streaming video editing techniques struggle to maintain consistent backgrounds and unedited areas while also achieving the low latency needed for a responsive user experience. Current methods designed for generating videos can’t directly be adapted for editing because they don’t reliably preserve existing content or allow precise control over specific regions within the video.

Listen to this article.

Problem

LLM agents are increasingly being used to tackle complex tasks, often involving multiple steps and interactions with external tools like web browsers or terminals. However, not every task is well-defined or even solvable within the available environment. This paper addresses a critical but largely overlooked problem: how do these agents decide when not to act – specifically, when to abstain from further action because continued attempts are unlikely to yield results? The authors term this “Agentic Abstention.” Current evaluation of LLM abstention often focuses on single-turn decisions; this work looks at the sequential decision making over multiple interactions.

Listen to this article.

Problem

Robotic manipulation often relies on simulated environments to train robots before deploying them in the real world. Current video generation models, even those fine-tuned for robotic tasks, struggle with physical plausibility. They frequently generate unrealistic movements and interactions, like objects bending unexpectedly or robot actions not making sense in a physics context. This lack of realism limits their usefulness as reliable world simulators for robot training.

Listen to this article.

Problem

Generating samples from molecular systems at thermodynamic equilibrium is computationally expensive and represents a significant hurdle in statistical physics. Current methods, known as Boltzmann Generators (BGs), attempt to speed up this process by combining generative models with precise likelihood calculations and importance sampling. However, existing BGs largely rely on normalizing flows, which have limitations – either expressing limited complexity or demanding computationally intensive operations.

Listen to this article.

Problem

Reinforcement learning (RL) has shown promise in improving large language models (LLMs). However, current RL methods often rely on having “ground-truth” answers to accurately reward the LLM’s performance. This severely limits their usefulness in situations where such ground truth is unavailable – a common scenario when dealing with tasks that involve complex problem-solving or code generation.

Method

The paper introduces a framework called RiVER (Ranking-induced VERifiable). The key innovation here is training LLMs on “score-based optimization tasks” rather than requiring ground-truth solutions. This means the model learns to improve based on execution feedback, specifically using scores as rewards – without needing to know the perfect answer upfront. The authors identified two issues when applying this approach: scale dominance (where different scores are skewed) and frequency dominance (where frequently sampled weaker solutions dominate learning). RiVER tackles these with a technique called “calibrated reward shaping” which uses comparisons between instances, emphasizing high-scoring solutions while still providing feedback for other valid results.

Listen to this article.

Problem

Training image generation models that excel at multiple tasks – like generating images from text (T2I), making local edits to existing images, and performing larger-scale global changes – is proving difficult. The authors of this paper point out a common issue: improving one capability often hurts another. For example, refining editing tools might reduce the quality of T2I generation, and trying to combine both local and global edits can lead to unexpected results.

Listen to this article.

Problem

Large language model (LLM) agents are increasingly relying on memory systems to store and retrieve information, evolving far beyond simple retrieval augmentation. However, current evaluations of these memory systems primarily focus on whether the agent succeeds in a task (using metrics like F1 score or BLEU). This overlooks crucial system-level considerations like cost, how different memory components work together, and how reliably the system handles knowledge updates over time – essentially treating everything as a black box.

Listen to this article.

Problem

Building truly general AI agents – systems that can effectively navigate and act in diverse, real-world environments – remains a significant challenge. A key component missing for these agents is a robust “world model”: the ability to predict how an environment will change based on actions taken within it. Current approaches struggle with accurately simulating agentic environments (where an actor interacts with the world).

Listen to this article.

Problem

Large language model (LLM) agents are being deployed to tackle increasingly complex, real-world tasks. These tasks often involve interacting with numerous tools – think of navigating a retail environment and needing to use various APIs or functions to find products, manage orders, track shipments, etc. Existing benchmarks haven’t adequately tested these agents’ ability to effectively plan across long sequences of tool usage, especially when dealing with limited visibility into which tools are available and reliable at any given moment.