Brief

Researchers have explored using off-the-shelf persona vectors to mitigate sycophancy in AI models, where models agree with users even when incorrect. They found that steering models towards personas exhibiting doubt or scrutiny significantly reduced sycophancy, performing comparably to methods specifically trained to combat this issue. Notably, this persona-based approach maintained model accuracy when users were correct, unlike traditional methods, and suggests sycophancy is more of a persona-level trait than a single steerable direction. AI

Researchers have developed a new framework called DABS for multi-aspect sentiment analysis, which aims to improve efficiency without sacrificing expressiveness. DABS encodes sentences only once, creating a reusable representation that aspects can query to selectively extract relevant information. This approach reduces computational costs by up to 60% in complex multi-aspect scenarios, particularly benefiting analyses involving negation and contrast. AI

Researchers have developed a hybrid machine learning model that integrates optical Landsat data with existing TanDEM-X interferometric measurements to improve forest height estimation. This enhanced model addresses ambiguities in previous methods by incorporating complementary information about forest type and structure. Validation against airborne LiDAR data showed a significant reduction in error, confirming the benefit of using multispectral inputs for more accurate remote sensing of forest parameters. AI

A Reddit user reported that the Cursor IDE consumed an unexpectedly large amount of memory, displaying a message indicating it was using gigabytes of RAM. The user expressed surprise at the high memory usage, noting that only three windows were open at the time. AI

A new research paper proposes a unified evidentiary framework for generative AI, combining cryptographic provenance, statistical watermarking, and zero-knowledge attestation. This framework aims to address legal challenges across international operational law, domestic court procedures, and product regulation. The study includes a benchmark of 12,000 generated items across various modalities and laundering pipelines, evaluating detection schemes and translating empirical bounds into legal sufficiency thresholds for different regulatory regimes. AI

Researchers have developed a new generative framework to model temporal processes in single-cell RNA sequencing data. This approach utilizes a latent heteroscedastic Gaussian process, approximated via Hilbert space methods, to capture population trends. An optimal transport objective is employed to align generated and observed distributions, addressing the challenge of inferring trajectories from static data. The method explicitly models biological heterogeneity by considering cell-specific latent time and cell type conditioning, demonstrating state-of-the-art performance on interpolation and extrapolation benchmarks. AI

A new research paper introduces DODOCO, a tool designed to diagnose overhead in dispatch operations for Mixture-of-Experts (MoE) models. The study found that common assumptions about workload characteristics and the effectiveness of existing mitigation strategies do not hold true for production routing. Specifically, the research indicates that scaling expert parallelism has minimal impact on routing imbalance, and mock-token benchmarks overestimate routing disparities compared to real text data. AI

Researchers have developed a new framework called PEACH that uses point clouds to adapt learned physics simulators to new material properties without needing explicit mesh reconstruction. This approach leverages in-context learning on point cloud sequences, improving simulation fidelity through novel encoding and auxiliary supervision. PEACH demonstrates accurate zero-shot sim-to-real transfer and outperforms mesh-based methods in prediction accuracy, making it more practical for real-world applications. AI

Researchers have introduced ArPoMeme, a new dataset containing approximately 7,300 Arabic political memes. This dataset is annotated with ideological orientations such as Leftist, Islamist, Pan-Arabist, and Satirical, as well as dimensions of polarization like Us vs. Them framing and hostility. The creation of ArPoMeme involved a semi-automated pipeline using web scraping and the Qwen2.5-VL-7B vision-language model for text extraction, followed by manual annotation via a custom interface. Analysis of the dataset indicates that Islamist and satirical memes exhibit the highest levels of hostility and mobilization cues. AI

Researchers have developed DrawMotion, a diffusion-based framework for generating 3D human motions that incorporates both text and hand-drawn sketches as input conditions. This dual-condition approach allows for more precise control over motion generation, with the hand-drawn element providing spatial guidance. Experiments show that using freehand drawings can reduce the time required for motion generation by nearly half compared to text-only methods. AI

A new service called RhymeRight is being launched, which will charge users $99 per month to determine if their sentences rhyme. The service is presented as a business idea with potential applications in creating bad poetry or confusing emails, framed humorously as a path to profit. AI

Researchers have developed a novel hybrid approach to estimate wheat spike volume using a combination of 3D reconstruction and knowledge distillation techniques. This method aims to overcome the challenges of traditional measurement methods, which are either computationally expensive or sensitive to environmental conditions. By distilling knowledge from a 3D model into a 2D image-based Transformer, the system achieves a significant reduction in mean absolute error and inference time, making it suitable for high-throughput field phenotyping. AI

Researchers have developed a new method called InterRS to enable AI to generate speech while simultaneously performing complex reasoning, mimicking human communication. This approach precisely interleaves reasoning steps within natural speech flow, requiring specially aligned data and a novel training pipeline. The method improves performance on logic and math benchmarks by 13% and produces more natural, fluent responses compared to existing techniques. AI

Researchers have introduced PaintCopilot, a novel AI system designed to assist in artistic painting by modeling the creative process as an autonomous continuation of prior artistic actions. Unlike methods that aim to reconstruct a target image, PaintCopilot generates future brushstrokes based on learned artistic dynamics and the evolving state of the canvas. The system comprises three models that predict artist intent, generate temporally coherent strokes, and synthesize localized sequences, enabling fluid co-creative workflows where artists and AI alternate control. AI

Aluminum recycling startups are increasingly leveraging artificial intelligence to improve their operations and capitalize on rising aluminum prices. These companies are integrating AI technologies to enhance sorting accuracy, optimize processing efficiency, and ultimately increase the yield of recycled aluminum. This strategic adoption of AI aims to make recycling more economically viable and environmentally sustainable. AI

Researchers have developed a new framework called the Combined Road Substrate (CRS) to improve visual reasoning for autonomous driving. CRS integrates geometric road structure with open-vocabulary semantics, allowing for more precise road understanding than current vision-language models. Training smaller models with CRS-enriched scenes significantly enhances their compositional reasoning abilities, shifting failure modes from relational understanding to attribute recognition, indicating that structured supervision is key rather than just model scale. AI

Researchers have developed DASH, a novel differentiable architecture search framework designed to rapidly discover efficient hybrid attention mechanisms for large language models. Unlike previous methods that required extensive computational resources, DASH significantly reduces search time and token usage by relaxing discrete operator placement into continuous logits and freezing model weights. This approach consistently yields superior results compared to existing baselines and even surpasses some released models, demonstrating that high-quality hybrid attention architectures can be found in minutes on a single GPU. AI

New York City Mayor Zohran Mamdani is launching a new Twitch show called "Talk with the People," set to premiere on May 21st. The show aims to engage with constituents by answering questions directly from the live chat about local issues. Mamdani plans to stream the series across multiple platforms, including YouTube and Facebook, to maximize reach. AI

Researchers have introduced the Winfree Oscillatory Neural Network (WONN), a novel dynamical architecture that leverages generalized Winfree dynamics for computation. This model represents data on a torus through structured oscillatory interactions, combining phase-based inductive biases with flexible interaction mechanisms. WONN has demonstrated competitive performance on image recognition and complex reasoning tasks, including ImageNet and Sudoku, while showing significant parameter efficiency compared to existing models. AI

Researchers have developed Strategy-Induct, a new framework for generating effective task-level instructions for large language models. This method bypasses the need for labeled answers by first prompting the model to create reasoning strategies for example questions. These strategy-question pairs are then used to induce a task instruction, which has shown superior performance compared to existing question-only approaches on various tasks and model scales. AI

Researchers have developed a new method to evaluate speech articulation synthesis by using phoneme recognition as a proxy for quality. This approach hypothesizes that articulatory features better capture phonetic nuances than traditional metrics. A neural network trained on acoustic and articulatory features from an RT-MRI dataset demonstrated that the proposed feature set is phonetically rich and aids in exploring new dimensions of speech articulation synthesis. AI

A new programming language called Sutra has been developed, designed to compile entire programs into fused tensor-operation graphs for PyTorch. This language targets Vector Symbolic Architectures and can represent complex logic, including Kleene connectives, as tensor operations. Sutra has demonstrated 100% accuracy in decoding bundles across various text and protein embeddings, outperforming standard Hadamard products, and its compiled graphs are fully differentiable, allowing for training and recompilation of the symbolic code. AI

Researchers have developed a new framework for implicit sentiment analysis, a task that infers sentiment from context rather than explicit words. Their approach, inspired by cognitive appraisal theory, uses a multi-task learning framework with two auxiliary tasks: implicit sentiment detection and cognitive rationale generation. To mitigate task interference, they implemented a task-routed mixture-of-experts model where tasks sparsely combine shared experts, outperforming existing methods on implicit sentiment tasks. AI

Researchers have developed a new reinforcement learning framework for fighting games that allows agents to learn not only which action to take but also for how long to execute it. This approach enables agents to dynamically adjust their responsiveness, moving beyond fixed decision-making intervals. Experiments in the FightLadder environment showed that learned timing can match fixed frame skips, but agents often performed best with higher frame skips, favoring exploitative strategies against scripted bots. AI

Researchers have developed new parallel and monolingual corpora specifically for scientific machine translation. These corpora focus on Spanish-English, French-English, and Portuguese-English language pairs, with specialized subsets for Cancer Research, Energy Research, Neuroscience, and Transportation. The created datasets were used to fine-tune general-purpose neural machine translation systems, and the paper details the corpus creation, fine-tuning methods, and evaluation results. AI

Researchers have developed a new framework called SynCB, which integrates concept-based models with standard neural networks. This hybrid approach uses a trainable routing module to dynamically select between a concept-based branch for interpretability and a complementary neural branch for performance. The two branches are learned jointly, allowing for information sharing and improved responsiveness to human interventions during testing. SynCB has demonstrated superior accuracy and intervention performance across multiple datasets compared to existing methods. AI

This paper delves into the computational complexity of entailment within cumulative propositional dependence logics and team semantics. It builds upon recent work characterizing these logics by System C and cumulative models, which allows for the analysis of entailment through relational models. AI

Microsoft has developed a Dynamic Threat Detection Agent (DTDA) integrated into its Security Copilot, designed to autonomously investigate security incidents and generate new detection logic. This agent utilizes a unified timeline of security data, LLM prompt contracts, and a planner-executor loop to identify hidden threats. In evaluations, DTDA achieved 80.1% precision and generated novel alerts for about 15% of investigated incidents, demonstrating its capability to find missed malicious activity at scale. AI

Researchers have developed VISTA, a system designed to anticipate human-object interactions in egocentric videos. VISTA combines spatial object detection with temporal context from video clips to predict future interactions, including object location, action categories, and timing. The system achieved first place in the EgoVis 2026 Ego4D Short-Term Object Interaction Anticipation Challenge. AI

Researchers have developed a new framework called HDMoE to improve multimodal cancer survival prediction. This hierarchical decoupling-fusion mixture-of-experts approach aims to better integrate data from sources like whole slide images and genomic profiles. The framework addresses limitations in existing methods by reducing redundant information before feature decoupling and by modeling fine-grained relationships within and between modalities. AI

A new research paper published on arXiv suggests that the current methods for predicting cancer drug sensitivity are flawed. The standard benchmark metric, global Pearson r, is misleading because it is heavily influenced by differences in drug potency rather than a model's ability to predict sensitivity for a specific tumor. When a more appropriate metric, per-drug Pearson r, is used, current drug encoding methods show no improvement over cell-only features. The study proposes that stratifying training data by mechanism-of-action can significantly improve prediction accuracy for targeted kinase inhibitors. AI

Researchers have developed a new framework called Map-Mono-Ego that enables accurate global human pose estimation using only a monocular camera. This method addresses the challenge of determining a user's absolute location within an environment, which is often overlooked by existing techniques that focus on relative motion. By integrating a pre-scanned 3D point cloud, Map-Mono-Ego overcomes the scale ambiguity inherent in monocular vision, preventing translational drift and enabling long-term tracking without specialized multi-sensor hardware. The effectiveness of this approach is further supported by the introduction of the AIST-Living dataset, which pairs egocentric video with ground-truth motion data in a scanned environment. AI

Researchers have developed a new dictionary learning method for fMRI data that accounts for individual brain geometry variations. This approach utilizes the optimal transport-based Fused Gromov-Wasserstein (FGW) distance to compare graphs with differing structures and features. To manage computational costs, they employ amortized optimization with a neural network to approximate optimal transport plans, enabling the learning of dictionary atoms that balance feature alignment and structural consistency. Experiments on the HCP dataset show this method effectively captures geometric variability and retains crucial information. AI

Researchers have introduced ProCrit, a novel framework for detecting multimodal sarcasm by employing a two-agent system. This system includes a proposal agent that generates diverse analytical perspectives and a critic agent that evaluates and guides revisions. To address the lack of detailed reasoning data, ProCrit synthesizes process-level annotations using a dynamic-role agentic rollout, creating sequences that preserve cross-perspective dependencies. The framework then refines both agents through a dual-stage reinforcement learning process, demonstrating effectiveness on multiple benchmarks. AI

Researchers have introduced Conditional Information Gain (CIG), a novel reward mechanism for reinforcement learning designed to improve exploration strategies. CIG addresses limitations of existing methods by providing a tractable surrogate for trajectory-level information gain, allowing it to scale to high-dimensional state spaces. Tested across twelve tasks in both discrete and continuous control environments, CIG demonstrated competitive or superior performance compared to previous exploration techniques, even in the presence of stochastic distractors. AI

Researchers have developed a policy system called CUGA designed to provide governance for generalist AI agents operating in enterprise environments. This system acts as a modular, policy-as-code layer that integrates with existing LLM agents without requiring model fine-tuning. CUGA enforces governance through five checkpoints: intent guarding, steering reasoning via playbooks, enforcing tool usage, human-in-the-loop approvals for risky actions, and output formatting. The system aims to ensure predictable, auditable, and compliance-aware behavior in complex workflows, as demonstrated in a healthcare scenario. AI

Researchers have developed CAdam, a new framework for generative distillation in 3D Gaussian Splatting that addresses limitations in adaptive densification. CAdam reinterprets densification as a signal verification problem, using gradient moments to distinguish consistent geometric signals from generative noise. This approach significantly reduces the number of Gaussian primitives needed while maintaining perceptual quality, improving memory efficiency in generative 3D tasks. AI