Each year, MIT Technology Review honors 35 outstanding people under the age of 35 who are driving scientific progress and solving tough problems in their fields. Today we want to introduce you to the computing innovators on the list who are coming up with new AI chips and specialized datasets--along with smart ideas about how to assess advanced systems for safety. Earlier this year, the $800 million Vera Rubin Observatory commenced its decade-long quest to create an extremely detailed time-lapse movie of the universe. Rubin is capable of capturing many more stars than any other astronomical observatory ever built; it also sees many more satellites. Up to 40% of images captured by the observatory within its first 10 years of operation will be marred by their sunlight-reflecting streaks. Meredith Rawls, a research scientist at the telescope's flagship observation project, Vera Rubin's Legacy Survey of Space and Time, is one of the experts tasked with protecting Rubin's science mission from the satellite blight.

The company behind ChatGPT is putting together a team capable of developing algorithms to control robots and appears to be hiring roboticists who work specifically on humanoids. OpenAI appears to be ramping up its efforts in robotics, hiring researchers who work on humanoid systems as it explores new ways to advance artificial intelligence . The company has recently recruited a number of researchers with expertise in developing AI algorithms for controlling humanoid and other types of robots. Job listings show that the company is putting together a team capable of creating systems that can be trained through teleoperation and simulation. Sources with knowledge of the company's efforts also say OpenAI is recruiting people to work specifically on humanoid robots, or robots with a partial or full human form.

North Korean hackers used ChatGPT to craft a fake draft of a South Korean military identification card to create a realistic-looking image meant to make a phishing attempt seem more credible, a South Korean cybersecurity firm found. A suspected North Korean state-sponsored hacking group used ChatGPT to create a deepfake of a military ID document to attack a target in South Korea, according to cybersecurity researchers. Attackers used the artificial intelligence tool to craft a fake draft of a South Korean military identification card in order to create a realistic-looking image meant to make a phishing attempt seem more credible, according to research published Sunday by Genians, a South Korean cybersecurity firm. Instead of including a real image, the email linked to malware capable of extracting data from recipients' devices, according to Genians. The group responsible for the attack, which researchers have dubbed Kimsuky, is a suspected North Korea-sponsored cyber-espionage unit previously linked to other spying efforts against South Korean targets.

Modern machine learning often requires training with large batch size, distributed data, and massively parallel compute hardware (like mobile and other edge devices or distributed data centers). Communication becomes a major bottleneck in such settings but methods like Local Stochastic Gradient Descent (Local SGD) show great promise in reducing this additional communication overhead. Local SGD consists of three parts: a local optimization process, an aggregation mechanism, and an outer optimizer that uses the aggregated updates from the nodes to produce a new model. While there exists an extensive literature on understanding the impact of hyperparameters in the local optimization process, the choice of outer optimizer and its hyperparameters is less clear. We study the role of the outer optimizer in Local SGD, and prove new convergence guarantees for the algorithm. In particular, we show that tuning the outer learning rate allows us to (a) trade off between optimization error and stochastic gradient noise variance, and (b) make up for ill-tuning of the inner learning rate. Our theory suggests that the outer learning rate should sometimes be set to values greater than $1$. We extend our results to settings where we use momentum in the outer optimizer, and we show a similar role for the momentum-adjusted outer learning rate. We also study acceleration in the outer optimizer and show that it improves the convergence rate as a function of the number of communication rounds, improving upon the convergence rate of prior algorithms that apply acceleration locally. Finally, we also introduce a novel data-dependent analysis of Local SGD that yields further insights on outer learning rate tuning. We conduct comprehensive experiments with standard language models and various outer optimizers to validate our theory.

Graph Neural Networks (GNNs) are widely used in learning on graph-structured data, yet a principled understanding of why they succeed or fail remains elusive. While prior works have examined architectural limitations such as over-smoothing and over-squashing, these do not explain what enables GNNs to extract meaningful representations or why performance varies drastically between similar architectures. These questions are related to the role of generalisation: the ability of a model to make accurate predictions on unlabelled data. Although several works have derived generalisation error bounds for GNNs, these are typically loose, restricted to a single architecture, and offer limited insight into what governs generalisation in practice. In this work, we take a different approach by deriving the exact generalisation error for GNNs in a transductive fixed-design setting through the lens of signal processing. From this viewpoint, GNNs can be interpreted as graph filter operators that act on node features via the graph structure. By focusing on linear GNNs while allowing non-linearity in the graph filters, we derive the first exact generalisation error for a broad range of GNNs, including convolutional, PageRank-based, and attention-based models. The exact characterisation of the generalisation error reveals that only the aligned information between node features and graph structure contributes to generalisation. Furthermore, we quantify the effect of homophily on generalisation. Our work provides a framework that explains when and why GNNs can effectively leverage structural and feature information, offering practical guidance for model selection.

Large-scale single-cell and Perturb-seq investigations routinely involve clustering cells and subsequently annotating each cluster with Gene-Ontology (GO) terms to elucidate the underlying biological programs. However, both stages, resolution selection and functional annotation, are inherently subjective, relying on heuristics and expert curation. We present HYPOGENEAGENT, a large language model (LLM)-driven framework, transforming cluster annotation into a quantitatively optimizable task. Initially, an LLM functioning as a gene-set analyst analyzes the content of each gene program or perturbation module and generates a ranked list of GO-based hypotheses, accompanied by calibrated confidence scores. Subsequently, we embed every predicted description with a sentence-embedding model, compute pair-wise cosine similarities, and let the agent referee panel score (i) the internal consistency of the predictions, high average similarity within the same cluster, termed intra-cluster agreement (ii) their external distinctiveness, low similarity between clusters, termed inter-cluster separation. These two quantities are combined to produce an agent-derived resolution score, which is maximized when clusters exhibit simultaneous coherence and mutual exclusivity. When applied to a public K562 CRISPRi Perturb-seq dataset as a preliminary test, our Resolution Score selects clustering granularities that exhibit alignment with known pathway compared to classical metrics such silhouette score, modularity score for gene functional enrichment summary. These findings establish LLM agents as objective adjudicators of cluster resolution and functional annotation, thereby paving the way for fully automated, context-aware interpretation pipelines in single-cell multi-omics studies.

This study addresses the escalating threat of branched broomrape (Phelipanche ramosa) to California's tomato industry, which supplies over 90 percent of U.S. processing tomatoes. The parasite's largely underground life cycle makes early detection difficult, while conventional chemical controls are costly, environmentally harmful, and often ineffective. To address this, we combined drone-based multispectral imagery with Long Short-Term Memory (LSTM) deep learning networks, using the Synthetic Minority Over-sampling Technique (SMOTE) to handle class imbalance. Research was conducted on a known broomrape-infested tomato farm in Woodland, Yolo County, CA, across five key growth stages determined by growing degree days (GDD). Multispectral images were processed to isolate tomato canopy reflectance. At 897 GDD, broomrape could be detected with 79.09 percent overall accuracy and 70.36 percent recall without integrating later stages. Incorporating sequential growth stages with LSTM improved detection substantially. The best-performing scenario, which integrated all growth stages with SMOTE augmentation, achieved 88.37 percent overall accuracy and 95.37 percent recall. These results demonstrate the strong potential of temporal multispectral analysis and LSTM networks for early broomrape detection. While further real-world data collection is needed for practical deployment, this study shows that UAV-based multispectral sensing coupled with deep learning could provide a powerful precision agriculture tool to reduce losses and improve sustainability in tomato production.

Large Language Models (LLMs) show promise in generating firmware for embedded systems, but often introduce security flaws and fail to meet real-time performance constraints. This paper proposes a three-phase methodology that combines LLM-based firmware generation with automated security validation and iterative refinement in a virtualized environment. Using structured prompts, models like GPT-4 generate firmware for networking and control tasks, deployed on FreeRTOS via QEMU. These implementations are tested using fuzzing, static analysis, and runtime monitoring to detect vulnerabilities such as buffer overflows (CWE-120), race conditions (CWE-362), and denial-of-service threats (CWE-400). Specialized AI agents for Threat Detection, Performance Optimization, and Compliance Verification collaborate to improve detection and remediation. Identified issues are categorized using CWE, then used to prompt targeted LLM-generated patches in an iterative loop. Experiments show a 92.4\% Vulnerability Remediation Rate (37.3\% improvement), 95.8\% Threat Model Compliance, and 0.87 Security Coverage Index. Real-time metrics include 8.6ms worst-case execution time and 195μs jitter. This process enhances firmware security and performance while contributing an open-source dataset for future research.

Alzheimer's disease is a progressive, neurodegenerative disorder that causes memory loss and cognitive decline. While there has been extensive research in applying deep learning models to Alzheimer's prediction tasks, these models remain limited by lack of available labeled data, poor generalization across datasets, and inflexibility to varying numbers of input scans and time intervals between scans. In this study, we adapt three state-of-the-art temporal self-supervised learning (SSL) approaches for 3D brain MRI analysis, and add novel extensions designed to handle variable-length inputs and learn robust spatial features. We aggregate four publicly available datasets comprising 3,161 patients for pre-training, and show the performance of our model across multiple Alzheimer's prediction tasks including diagnosis classification, conversion detection, and future conversion prediction. Importantly, our SSL model implemented with temporal order prediction and contrastive learning outperforms supervised learning on six out of seven downstream tasks. It demonstrates adaptability and generalizability across tasks and number of input images with varying time intervals, highlighting its capacity for robust performance across clinical applications.

Transformer-based language models are architecturally constrained to process text of a fixed maximum length. Essays written by higher-grade students frequently exceed the maximum allowed length for many popular open-source models. A common approach to addressing this issue when using these models for Automated Essay Scoring is to truncate the input text. This raises serious validity concerns as it undermines the model's ability to fully capture and evaluate organizational elements of the scoring rubric, which requires long contexts to assess. In this study, we evaluate several models that incorporate architectural modifications of the standard transformer architecture to overcome these length limitations using the Kaggle ASAP 2.0 dataset. The models considered in this study include fine-tuned versions of XLNet, Longformer, ModernBERT, Mamba, and Llama models.