AI Research & Papers — September 22, 2026

Findings

[1] 2026-09-22 Poitras Center to fuel early careers of 50 young scientists dedicated to psychiatric disorders research

Patricia and James Poitras ’63, longtime MIT supporters, have launched a fellowship program for graduate students and postdocs studying major mental illness, expanding their MIT philanthropy to directly support early-career scientists. The commitment establishes 50 two-year fellowships through the Poitras Center for Psychiatric Disorders Research at MIT’s McGovern Institute for Brain Research. Five fellowships will be awarded every year for the next decade,…
[2] 2026-09-22 How Do LLMs Cite? A Mechanistic Interpretation of Attribution in Retrieval-Augmented Generation

arXiv:2606.28358v2 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) aims to enhance the trustworthiness of Large Language Models (LLMs) by grounding their outputs in external documents, often using inline citations for verifiability. However, the faithfulness of these citations — whether the model genuinely uses a source to generate an answer — remains a critical, unverified assumption. This paper offers the first…
[3] 2026-09-22 Evaluating Fine-Tuned and Base Language Models in Maternal and Vaccination Healthcare for African Settings

arXiv:2609.22110v1 Announce Type: new Abstract: Background: Large language models (LLMs) can improve healthcare information delivery in low-resource settings but may produce inaccurate or culturally inappropriate advice. This study evaluated domain-specific fine-tuning for maternal health and vaccination in Nigeria. Objective: To compare HelpMum's MamaBot-Llama and Vax-Llama with Meta's Llama-3.1-8B-Instruct for accuracy, safety, clarity, contextual appropriateness, and trustworthiness. Methods: We evaluated 200…
[4] 2026-09-22 A frontend-backend architecture for tool calls in full-duplex speech models

arXiv:2609.19334v2 Announce Type: replace Abstract: Full-duplex speech-to-speech (S2S) models provide natural, low-latency conversational interaction and would benefit from the ability to use external tools and complete voice-agent tasks. We propose a frontend-backend architecture where a duplex speech-to-text frontend learns to emit a delegation token and forwards streaming ASR transcripts to a text-based backend LLM for tool calls. Tool-call results from…
[5] 2026-09-22 DeepInstructor: An Agentic AI Instructor for Experience-Driven Idea Evaluation

arXiv:2609.22104v1 Announce Type: new Abstract: As automated scientific discovery advances, Large Language Models (LLMs) can now generate research ideas at an unprecedented scale, shifting the bottleneck from idea generation to idea evaluation. Existing evaluators mainly rely on parametric LLM knowledge or unstructured retrieval, producing judgments that lack the experience-grounded reasoning used by human instructors. To address this, we propose DeepInstructor,…
[6] 2026-09-22 The Harder Text Embedding Benchmark (HTEB): Beyond One-dimensional Static Robustness

arXiv:2605.28190v2 Announce Type: replace Abstract: Embedding benchmarks like MTEB report a single score per model, implicitly treating robustness as a static, scalar property. We argue that embedding robustness is multidimensional, since models respond differently to different types of variation, and requires dynamic evaluation to expose failures hidden by static benchmarks. We introduce the Harder Text Embedding Benchmark (HTEB), a dynamic…
[7] 2026-09-22 Context Poisoning as Extreme-Value Attention Interference in Long-Context Language Models

arXiv:2609.22101v1 Announce Type: new Abstract: Large language models can process increasingly long prompts, yet their ability to locate and use decisive evidence may degrade as irrelevant or confusable context is added. We formulate this phenomenon, which we call context poisoning, as extreme-value interference in attention: the decisive-evidence score is upper-bounded, while the maximum score among effective distractors grows with their…
[8] 2026-09-22 Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models

arXiv:2602.01654v2 Announce Type: replace Abstract: Steering vectors (SVs) offer a lightweight way to control large language models (LLMs) at inference time by shifting hidden activations, providing a practical middle ground between prompting and fine-tuning. Yet SVs can be unreliable in practice. Some concepts are unsteerable, and even when steering helps on average it can backfire for a non-trivial fraction of…
[9] 2026-09-22 Summarize, Judge, Refine: Decoupled Content Understanding and Policy Learning for Multimodal Content Moderation

arXiv:2609.22094v1 Announce Type: new Abstract: Content moderation systems traditionally entangle multimodal understanding with policy-specific classification, requiring full pipeline retraining for every policy change and suffering from label scarcity since multimedia cannot be meaningfully augmented. We propose Summarize-Judge-Refine (SJR), a two-model architecture that decouples these concerns via a natural language interface: a multimodal Content Model produces structured text summaries, and a…
[10] 2026-09-22 AI-inferred expressed well-being and collective-action discourse in climate-change campaigns on X

arXiv:2609.22096v1 Announce Type: new Abstract: Climate campaigns are often evaluated through attention and mobilization, but less is known about the well-being language that accompanies them. Whether campaign periods alter positive affect and hope, and whether happiness aligns with action language, remains unresolved. We analysed 364,118 public Twitter/X posts from Earth Day, Earth Hour, Global Climate Action Day and World Environment…
[11] 2026-09-22 Token Signatures of Code: Comparing Coding Behaviors Across Large Language Models

arXiv:2609.22097v1 Announce Type: new Abstract: The evaluation of large language models (LLMs) on coding tasks has primarily focused on performance metrics such as pass@k. As LLMs continue to advance, many models now meet baseline performance requirements, reducing the discriminative power of performance-based evaluation alone. Yet a key question remains largely unexplored: how do LLMs differ in their coding behavior? We…
[12] 2026-09-22 A framework for recipe data structure with applications for culinary and nutritional insights

arXiv:2609.22099v1 Announce Type: new Abstract: Cooking is a complex process that transforms raw ingredients into delicious and nutritious dishes, yet the recipes that encode this process remain largely free text; readable by people but not directly computable. Existing recipe collections capture fragments of this information, but no shared representation links a recipe's structured ingredient composition, its geo-cultural provenance, and its…
[13] 2026-09-22 TreeSpark: Calibrated, Load-Adaptive Draft Trees for Semi-Autoregressive Speculative Decoding

arXiv:2609.22098v1 Announce Type: new Abstract: Speculative decoding accelerates language-model inference by letting a cheap drafter propose tokens that the target model verifies in parallel. Recent block drafters make drafting nearly free: a single backbone pass emits an entire block of draft tokens. Draft trees promise a further gain — several alternative continuations verified in one target forward — but existing…
[14] 2026-09-22 Counting Documents Is Not Counting Text: Unit Bias in Web-PDF Corpus Statistics

arXiv:2608.16390v2 Announce Type: replace Abstract: PDF corpora advertise their size in tokens, but every rate they publish (coverage, OCR routing, re-fetch recovery, language mix) is computed per document, and none decomposes its token total. Because PDF length is extremely skewed, the two units can describe the same corpus very differently. We ask how the headline statistics of a web-PDF corpus…
[15] 2026-09-22 AdaMem: Adaptive Memory Token Allocation for Soft Compression in Retrieval-Augmented Generation

arXiv:2609.22100v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) improves language models with retrieved evidence, but processing many long passages is costly and can introduce distracting information. Soft compression addresses this challenge by encoding passages as compact sequences of continuous memory embeddings before generation. However, existing methods typically assign each retained passage an identical number of memory embeddings, irrespective of its…
[16] 2026-09-22 CovR: Coverage-Aware Hardware Verification via Reasoning-Guided Reinforcement Learning

arXiv:2609.19189v2 Announce Type: replace-cross Abstract: Design verification remains one of the most resource-intensive stages of hardware development, often consuming up to 70% of the total design effort. While recent work has explored using Large Language Models (LLMs) to automate testbench generation, most existing approaches focus narrowly on functional correctness, overlooking the critical aspect of coverage quality. To bridge this gap,…
[17] 2026-09-22 A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agents

arXiv:2601.12538v2 Announce Type: replace-cross Abstract: Reasoning is a fundamental cognitive process underlying inference, problem-solving, and decision-making. While large language models (LLMs) demonstrate strong reasoning capabilities in closed-world settings, they struggle in open-ended and dynamic environments. Agentic reasoning marks a paradigm shift by reframing LLMs as autonomous agents that plan, act, and learn through continual interaction. In this survey, we organize…
[18] 2026-09-22 MuLA-Bench: A Multilingual Long-Form Audio Understanding Benchmark via Multi-Tier Auditing

arXiv:2609.23416v1 Announce Type: cross Abstract: Long-form audio performance is often summarized by context length and aggregate accuracy, obscuring how language, evidence, and task jointly shape difficulty. We introduce MuLA-Bench: 5,038 open-ended questions over 1,769 in-the-wild recordings totaling 1,377.9 hours, covering 16 languages and eight domains. A balanced Language x Domain semantic track supports controlled comparisons, while a complementary acoustic track…
[19] 2026-09-22 Strategy Accumulation and Guided Execution for Automated LLM Fine-Tuning

arXiv:2609.22257v1 Announce Type: cross Abstract: Producing task-specific large language models requires discovering effective training strategies through experimentation. Automated fine-tuning systems have made this experimentation feasible with far less manual effort. However, these systems are stateless: each search discards its discovered strategies, dataset insights, and hyperparameter findings once it ends. Every new task must then repeat this costly search from a…
[20] 2026-09-22 Compound-QA: A Benchmark for Evaluating LLMs on Compound Questions

arXiv:2411.10163v3 Announce Type: replace Abstract: Large language models (LLMs) demonstrate remarkable performance across various tasks, prompting researchers to develop diverse evaluation benchmarks. However, most benchmarks typically measure the ability of LLMs to respond to individual questions, neglecting the complex interactions in real-world applications. We introduce Compound Question Synthesis (CQ-Syn) to build Compound-QA, a benchmark targeting questions composed of multiple interrelated…
[21] 2026-09-22 UK-PRBENCH: A Paragraph-Level Precedent Retrieval Benchmark for United Kingdom Case Law

arXiv:2609.24613v1 Announce Type: cross Abstract: Prior case retrieval (PCR) aims to identify precedent cases relevant to a given query case. Existing PCR benchmarks and methods predominantly operate at the document level, treating entire judgments as the unit of relevance. This formulation is suboptimal for legal practitioners, as judgments address multiple legal issues and only a small subset of paragraphs is…
[22] 2026-09-22 SolarFlowRefiner: Refinement-Aware Flow Matching for Surface Solar Radiation Downscaling

arXiv:2609.22126v1 Announce Type: new Abstract: High-resolution surface solar radiation (SSR) is important for solar forecasting and grid operation. However, physically consistent reanalysis products are too coarse to resolve localized cloud-driven variability. In this paper, we study a multisource downscaling task that reconstructs high-resolution SolarCube SSR fields from coarse ERA5 radiative variables and co-registered satellite channels. The task is challenging because…
[23] 2026-09-22 $N_0$-Foundation: Towards the Age of Tactile Intelligence

arXiv:2608.29601v2 Announce Type: replace-cross Abstract: We present $N_0$-Foundation, a paradigm for tactile-enabled embodied manipulation, which integrates tactile sensing hardware, large-scale multimodal data, tactile representation learning, and standardized evaluation. First, we engineer the infrastructure for scalable data collection, including a vision-based tactile sensor, a tactile Universal Manipulation Interface (UMI), and a synchronized visuo-tactile data collection system supporting both robot embodiments and…
[24] 2026-09-22 StationPDE: Station-Oriented Surface PDE Learning for Multi-Station Multivariate Weather Forecasting

arXiv:2609.22123v1 Announce Type: new Abstract: Multi-station multivariate weather forecasting aims to forecast future weather variables at multiple weather stations from historical surface observations. Existing station forecasting models learn statistical dependencies among discrete stations, but lack explicit physical evolution. Meanwhile, PDE-based weather models provide interpretable physical dynamics, yet require continuous fields and upper-air variables unavailable in surface station data. To bridge…
[25] 2026-09-22 Explain Less, Understand More: Data-Efficient Personalization of Reader-Dependent Jargons

arXiv:2505.16227v4 Announce Type: replace-cross Abstract: Personalizing jargon detection and explanation is essential for making technical documents accessible to readers with diverse disciplinary backgrounds. However, tailoring models to individual users typically requires substantial annotation efforts and computational resources due to user-specific finetuning. To address this, we present a systematic study of personalized jargon detection, focusing on methods that are both efficient…
[26] 2026-09-22 Rank Portability Does Not Imply Feasibility Portability: Target-Specific Evaluation of Joint Hardware Constraints

arXiv:2609.22122v1 Announce Type: new Abstract: Cross-device hardware evaluation often assumes that if architecture rankings transfer across devices, a proxy device can support target-side model selection. We stress-test this assumption for joint latency-energy feasibility across two public architecture families. On NAS-Bench-201, cross-device rank correlations are moderate, while target-comparable feasible-set overlap remains incomplete. A faithful AdaProxy diagnostic substantially improves latency ranking, showing…
[27] 2026-09-22 Robust small-molecule identification from incomplete, degraded, and inconsistent spectra using multimodal mixed-condition training

arXiv:2609.14360v2 Announce Type: replace Abstract: Reliable small-molecule identification often requires complementary evidence from multiple spectroscopic measurements. In practice, however, spectra may be unavailable, degraded by measurement-related variations, or even incorrectly associated with a sample, thereby hindering accurate molecular identification. Herein, we propose a multimodal mixed-condition training strategy that accommodates missing, degraded, and mismatched measurements for small-molecule structure identification. The strategy…
[28] 2026-09-22 Correcting Learning-based Perception for Safety

arXiv:2609.22108v1 Announce Type: new Abstract: Learning-enabled perception is important in many autonomous systems. Unlike traditional sensors, the boundary where ML perception does or does not work is poorly characterized. Incorrect perception can lead to unsafe or overtly conservative downstream control actions. In this paper, we propose a two-step strategy for correcting ML-based state estimation. First, an offline computation is used…
[29] 2026-09-22 A Shared Learning Rate Is Not a Neutral Control in Selective On-Policy Distillation

arXiv:2609.22109v1 Announce Type: new Abstract: Selective on-policy distillation trains a student only at the token positions a selector scores highest, and the literature compares selectors under a single shared learning rate–a control chosen to be neutral. We show it is not. Under LoRA on GSM8K (Qwen2.5-1.5B student, 7B teacher), across an 8x learning-rate grid, dense supervision is statistically flat (swing…
[30] 2026-09-22 Toward Fairness in Machine Learning Models for Predicting Treatment Retention and Premature Discontinuation in Medication for Opioid Use Disorder

arXiv:2609.22113v1 Announce Type: new Abstract: Persistent low retention and completion rates in medications for opioid use disorder (MOUD) have driven the use of machine learning (ML) models to predict retention and identify patients at risk of premature discontinuation. However, the fairness of these models across patient populations remains largely unexplored, raising concerns about their application in treatment decision support. This…
[31] 2026-09-22 LE4Mob: Towards Inductive, Distance-Aware and General-Purpose Location Embedding for Human Mobility Modelling

arXiv:2609.22117v1 Announce Type: new Abstract: Location representations provide mobility models with fundamental information about the spatial position, functional characteristics, and relationships of places. However, existing embeddings are often dependent on mobility observations, unable to represent unseen locations, and weakly constrained to retain geographic distance. This limits their reuse across datasets and mobility tasks. To address these limitations, we propose LE4Mob,…
[32] 2026-09-22 ZoAQ: Adaptive Zeroth-Order Querying via Query-Reuse Coupling

arXiv:2609.22115v1 Announce Type: new Abstract: Zeroth-order optimization (ZOO) estimates updates from function evaluations, making perturbation queries a primary cost. Fixed budgets spend the same number of queries at every step, while adaptive controllers may offset their savings by using additional oracle calls to test estimator reliability. We introduce ZoAQ, an adaptive ZOO method built around query reuse. Rather than discarding…
[33] 2026-09-22 Modelling daily activity patterns from mobile phone location data via deep representation learning

arXiv:2609.22121v1 Announce Type: new Abstract: Passively collected mobile phone location data provide large-scale, longitudinal observations of human mobility but do not directly reveal activity purposes. The functional characteristics of visited locations offer useful contextual information, yet their relationship with activity purpose remains uncertain, particularly in mixed-use urban environments. We conceptualise activity pattern mining as an integrated process of representation, clustering,…
[34] 2026-09-22 Success Leaves Detours: Learning Executable Walkthroughs for Long-Horizon Agents

arXiv:2609.22120v1 Announce Type: new Abstract: Test-time self-evolving agents improve by reusing past experience, yet sparse-reward trajectories contain failures, loops, and detours, while summaries often omit the state conditions and action dependencies needed for execution. We study executable Walkthrough induction from sparse-reward trajectories: extracting compact, state-conditioned, and verifiable procedures. Our key observation is that delayed credit identifies actions associated with progress…
[35] 2026-09-22 Scalable Mamba-Based Message-Passing Neural Decoder for Error-Correcting Codes

arXiv:2605.10681v2 Announce Type: replace-cross Abstract: Forward error correction is essential for reliable communication over noisy channels. Attention-based model-free neural decoders have shown strong performance for short codes, but their scalability to longer codes is limited by the quadratic memory and computational cost of attention. In this paper, we introduce the Mamba message-passing decoder (MMPD), an attention-free syndrome-based neural decoder for…
[36] 2026-09-22 The shape of quark flavors

arXiv:2609.22812v1 Announce Type: cross Abstract: We construct the Yukawa couplings of the quark sector as overlap integrals of Gaussian wave functions in extra spatial dimensions. Assuming that the wave function of each chiral fermion is localised at a point in the extra dimensions, while that of the Higgs doublet is flat, the Yukawa matrix elements are determined by the relative…
[37] 2026-09-22 Exponential Family Synthetic Controls

arXiv:2609.23970v1 Announce Type: cross Abstract: We develop exponential family synthetic controls (EFSC), a distributional version of synthetic controls for a panel of datasets. Each cell of the panel corresponds to a dataset drawn from an exponential family whose natural parameters factorize probabilistically across units and times. We estimate the latent factors using black-box variational inference. This replaces the usual weighted-average…
[38] 2026-09-22 Stochastic Reconfiguration as Statistical Filtering for Overparameterized Neural Quantum States

arXiv:2609.23334v1 Announce Type: cross Abstract: Stochastic reconfiguration (SR) is the standard optimizer for neural quantum states (NQS), but modern NQS often have far more parameters than Monte Carlo samples. We show that in this regime the diagonal shift is more than a numerical stabilizer. It acts as a statistical filter for finite-sample generalization. At a fixed wave function, SR is…
[39] 2026-09-22 Find Your Optimal Teacher: Personalized Data Synthesis via Router-Guided Multi-Teacher Distillation

arXiv:2510.10925v3 Announce Type: replace Abstract: Training student models on synthetic data generated by strong teacher models is a promising way to distilling the capabilities of teachers. However, recent studies show that stronger models are not always optimal teachers, revealing a mismatch between teacher outputs and student learnability. To address this issue, we propose PerSyn (Personalized data Synthesis), a novel synthesis…
[40] 2026-09-22 D-JEPA: A Decision-Aligned Latent World Model

arXiv:2609.24749v1 Announce Type: cross Abstract: Latent world models predict the consequences of actions, but accurate prediction does not guarantee that latent distance reflects which candidate will execute successfully. We identify a decision-local prediction gap: among the few futures competing for execution, a candidate predicted closer to the goal can produce a worse realized outcome than an available alternative. We introduce…
[41] 2026-09-22 The Illusion of Improvement: Reject Inference Strategies in Credit Scoring

arXiv:2606.18479v3 Announce Type: replace Abstract: Reject inference methods are widely used to mitigate survival bias in credit scoring, yet their effectiveness remains poorly understood. We systematically evaluate several such methods and uncover a structural failure mode: in a natural retraining cycle, models whose accuracy improves while recall collapses create an illusion of improvement that leads practitioners to believe the system…

Where Kimbodo Comes In

Kimbodo builds and operates this in production for businesses — see our AI Consulting & Strategy practice, or Request an AI Roadmap.

AI Research & Papers — September 22, 2026

Findings

Where Kimbodo Comes In

Sources

Chad Collins

Leave a comment Cancel reply

You May Also Like

How to Deploy Lower‑Cost, Better‑Grounded, Safer AI Systems Using Recent Research Breakthroughs

Which AI Research Should Business Leaders Prioritize: Trust, Long‑Memory, Retrieval, and Efficient Serving

Engineering Founded & Customer Focused