Search

Large language model (LLM)-based agents have shown promise in tackling complex tasks by interacting dynamically with the environment. Existing work primarily focuses on behavior cloning from expert demonstrations and preference learning through exploratory trajectory sampling. However, these methods often struggle in long-horizon tasks, where suboptimal actions accumulate step by step, causing agents to deviate from correct task trajectories. To address this, we highlight the importance of timely calibration and the need to automatically construct calibration trajectories for training agents. We propose Step-Level Trajectory Calibration (STeCa), a novel framework for LLM agent learning. Specifically, STeCa identifies suboptimal actions through a step-level reward comparison during exploration. It constructs calibrated trajectories using LLM-driven reflection, enabling agents to learn from improved decision-making processes. These calibrated trajectories, together with successful trajectory data, are utilized for reinforced training. Extensive experiments demonstrate that STeCa significantly outperforms existing methods. Further analysis highlights that step-level calibration enables agents to complete tasks with greater robustness. Our code and data are available at https://github.com/WangHanLinHenry/STeCa.

The safety alignment of large language models (LLMs) remains vulnerable, as their initial behavior can be easily jailbroken by even relatively simple attacks. Since infilling a fixed template between the input instruction and initial model output is a common practice for existing LLMs, we hypothesize that this template is a key factor behind their vulnerabilities: LLMs' safety-related decision-making overly relies on the aggregated information from the template region, which largely influences these models' safety behavior. We refer to this issue as template-anchored safety alignment. In this paper, we conduct extensive experiments and verify that template-anchored safety alignment is widespread across various aligned LLMs. Our mechanistic analyses demonstrate how it leads to models' susceptibility when encountering inference-time jailbreak attacks. Furthermore, we show that detaching safety mechanisms from the template region is promising in mitigating vulnerabilities to jailbreak attacks. We encourage future research to develop more robust safety alignment techniques that reduce reliance on the template region.

Chain-of-Thought (CoT) has been proven effective in enhancing the reasoning capabilities of large language models (LLMs). Recent advancements, such as OpenAI's o1 and DeepSeek-R1, suggest that scaling up the length of CoT sequences during inference could further boost LLM reasoning performance. However, due to the autoregressive nature of LLM decoding, longer CoT outputs lead to a linear increase in inference latency, adversely affecting user experience, particularly when the CoT exceeds 10,000 tokens. To address this limitation, we analyze the semantic importance of tokens within CoT outputs and reveal that their contributions to reasoning vary. Building on this insight, we propose TokenSkip, a simple yet effective approach that enables LLMs to selectively skip less important tokens, allowing for controllable CoT compression. Extensive experiments across various models and tasks demonstrate the effectiveness of TokenSkip in reducing CoT token usage while preserving strong reasoning performance. Notably, when applied to Qwen2.5-14B-Instruct, TokenSkip reduces reasoning tokens by 40% (from 313 to 181) on GSM8K, with less than a 0.4% performance drop.

Daily activity recognition has gained prominence due to its applications in context-aware computing. Current methods primarily rely on supervised learning for detecting simple, repetitive activities. This paper introduces LayeredSense, a novel framework designed to recognize complex activities by decomposing them into smaller, easily identifiable unit patterns. Utilizing a Myo armband for data collection, our system processes inertial measurement unit (IMU) data to identify basic actions like walking, running, and jumping. These actions are then aggregated to infer more intricate activities such as playing sports or working. LayeredSense employs Gaussian Mixture Models for new pattern detection and machine learning algorithms, including Random Forests, for real-time activity recognition. Our system demonstrates high accuracy in identifying both unit patterns and complex activities, providing a scalable solution for comprehensive daily activity monitoring

Background: RETFound, a self-supervised, retina-specific foundation model (FM), showed potential in downstream applications. However, its comparative performance with traditional deep learning (DL) models remains incompletely understood. This study aimed to evaluate RETFound against three ImageNet-pretrained supervised DL models (ResNet50, ViT-base, SwinV2) in detecting ocular and systemic diseases. Methods: We fine-tuned/trained RETFound and three DL models on full datasets, 50%, 20%, and fixed sample sizes (400, 200, 100 images, with half comprising disease cases; for each DR severity class, 100 and 50 cases were used. Fine-tuned models were tested internally using the SEED (53,090 images) and APTOS-2019 (3,672 images) datasets and externally validated on population-based (BES, CIEMS, SP2, UKBB) and open-source datasets (ODIR-5k, PAPILA, GAMMA, IDRiD, MESSIDOR-2). Model performance was compared using area under the receiver operating characteristic curve (AUC) and Z-tests with Bonferroni correction (P<0.05/3). Interpretation: Traditional DL models are mostly comparable to RETFound for ocular disease detection with large datasets. However, RETFound is superior in systemic disease detection with smaller datasets. These findings offer valuable insights into the respective merits and limitation of traditional models and FMs.

The o1 model series is trained with large-scale reinforcement learning to reason using chain of thought. These advanced reasoning capabilities provide new avenues for improving the safety and robustness of our models. In particular, our models can reason about our safety policies in context when responding to potentially unsafe prompts, through deliberative alignment. This leads to state-of-the-art performance on certain benchmarks for risks such as generating illicit advice, choosing stereotyped responses, and succumbing to known jailbreaks. Training models to incorporate a chain of thought before answering has the potential to unlock substantial benefits, while also increasing potential risks that stem from heightened intelligence. Our results underscore the need for building robust alignment methods, extensively stress-testing their efficacy, and maintaining meticulous risk management protocols. This report outlines the safety work carried out for the OpenAI o1 and OpenAI o1-mini models, including safety evaluations, external red teaming, and Preparedness Framework evaluations.

We consider the challenge of black-box optimization within hybrid discrete-continuous and variable-length spaces, a problem that arises in various applications, such as decision tree learning and symbolic regression. We propose DisCo-DSO (Discrete-Continuous Deep Symbolic Optimization), a novel approach that uses a generative model to learn a joint distribution over discrete and continuous design variables to sample new hybrid designs. In contrast to standard decoupled approaches, in which the discrete and continuous variables are optimized separately, our joint optimization approach uses fewer objective function evaluations, is robust against non-differentiable objectives, and learns from prior samples to guide the search, leading to significant improvement in performance and sample efficiency. Our experiments on a diverse set of optimization tasks demonstrate that the advantages of DisCo-DSO become increasingly evident as the complexity of the problem increases. In particular, we illustrate DisCo-DSO's superiority over the state-of-the-art methods for interpretable reinforcement learning with decision trees., Comment: Accepted at AAAI-25

Large Language Models (LLMs) have revolutionized text generation, making detecting machine-generated text increasingly challenging. Although past methods have achieved good performance on detecting pure machine-generated text, those detectors have poor performance on distinguishing machine-revised text (rewriting, expansion, and polishing), which can have only minor changes from its original human prompt. As the content of text may originate from human prompts, detecting machine-revised text often involves identifying distinctive machine styles, e.g., worded favored by LLMs. However, existing methods struggle to detect machine-style phrasing hidden within the content contributed by humans. We propose the "Imitate Before Detect" (ImBD) approach, which first imitates the machine-style token distribution, and then compares the distribution of the text to be tested with the machine-style distribution to determine whether the text has been machine-revised. To this end, we introduce style preference optimization (SPO), which aligns a scoring LLM model to the preference of text styles generated by machines. The aligned scoring model is then used to calculate the style-conditional probability curvature (Style-CPC), quantifying the log probability difference between the original and conditionally sampled texts for effective detection. We conduct extensive comparisons across various scenarios, encompassing text revisions by six LLMs, four distinct text domains, and three machine revision types. Compared to existing state-of-the-art methods, our method yields a 13% increase in AUC for detecting text revised by open-source LLMs, and improves performance by 5% and 19% for detecting GPT-3.5 and GPT-4o revised text, respectively. Notably, our method surpasses the commercially trained GPT-Zero with just $1,000$ samples and five minutes of SPO, demonstrating its efficiency and effectiveness., Comment: To appear at AAAI 2025. 14 pages, 6 figure

Investigating the signals of dark matter annihilation is one of the most popular ways to understand the nature of dark matter. In particular, many recent studies are focussing on using radio data to examine the possible signals of dark matter revealed in galaxies and galaxy clusters. In this article, we investigate on the spectral data of the central radio halo of the cool-core cluster RX J1720.1+2638. We show that the radio spectral data can be best accounted by the synchrotron emission due to dark matter annihilation via $\tau$ lepton channel (with dark matter mass $m=15$ GeV) or $b$ quark channel (with dark matter mass $m=110$ GeV), although using the very coarse spectral data with notable errors. Despite the fact that cosmic-ray emission can also provide a good explanation for the observed radio spectrum, our results suggest a possible positive evidence for dark matter annihilation revealed in the form of radio emission in RX J1720.1+2638 cluster., Comment: Accepted in MNRAS

Transcatheter tricuspid valve replacement (TTVR) is the latest treatment for tricuspid regurgitation and is in the early stages of clinical adoption. Intelligent robotic approaches are expected to overcome the challenges of surgical manipulation and widespread dissemination, but systems and protocols with high clinical utility have not yet been reported. In this study, we propose a complete solution that includes a passive stabilizer, robotic drive, detachable delivery catheter and valve manipulation mechanism. Working towards autonomy, a hybrid augmented intelligence approach based on reinforcement learning, Monte Carlo probabilistic maps and human-robot co-piloted control was introduced. Systematic tests in phantom and first-in-vivo animal experiments were performed to verify that the system design met the clinical requirement. Furthermore, the experimental results confirmed the advantages of co-piloted control over conventional master-slave control in terms of time efficiency, control efficiency, autonomy and stability of operation. In conclusion, this study provides a comprehensive pathway for robotic TTVR and, to our knowledge, completes the first animal study that not only successfully demonstrates the application of hybrid enhanced intelligence in interventional robotics, but also provides a solution with high application value for a cutting-edge procedure.

Existing guarantees for algorithms sampling from nonlogconcave measures on $\mathbb{R}^d$ are generally inexplicit or unscalable. Even for the class of measures with logdensities that have bounded Hessians and are strongly concave outside a Euclidean ball of radius $R$, no available theory is comprehensively satisfactory with respect to both $R$ and $d$. In this paper, it is shown that complete polynomial complexity can in fact be achieved if $R\leq c\sqrt{d}$, whilst an exponential number of point evaluations is generally necessary for any algorithm as soon as $R\geq C\sqrt{d}$ for constants $C>c>0$. A simple importance sampler with tail-matching proposal achieves the former, owing to a blessing of dimensionality. On the other hand, if strong concavity outside a ball is replaced by a distant dissipativity condition, then sampling guarantees must generally scale exponentially with $d$ in all parameter regimes.

The alignment of large language models (LLMs) with human preferences remains a key challenge. While post-training techniques like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) have achieved notable success, they often introduce computational inefficiencies and training instability. In this paper, we propose Feature-level constrained Preference Optimization (FPO), a novel method designed to simplify the alignment process while ensuring stability. FPO leverages pre-trained Sparse Autoencoders (SAEs) and introduces feature-level constraints, allowing for efficient, sparsity-enforced alignment. Our approach enjoys efficiency by using sparse features activated in a well-trained sparse autoencoder and the quality of sequential KL divergence by using the feature-level offline reference. Experimental results on benchmark datasets demonstrate that FPO achieves a 5.08% absolute improvement in win rate with much lower computational cost compared to state-of-the-art baselines, making it a promising solution for efficient and controllable LLM alignments.

Effectively modeling multimodal longitudinal data is a pressing need in various application areas, especially biomedicine. Despite this, few approaches exist in the literature for this problem, with most not adequately taking into account the multimodality of the data. In this study, we developed multiple configurations of a novel multimodal and longitudinal learning framework, Longitudinal Ensemble Integration (LEI), for sequential classification. We evaluated LEI's performance, and compared it against existing approaches, for the early detection of dementia, which is among the most studied multimodal sequential classification tasks. LEI outperformed these approaches due to its use of intermediate base predictions arising from the individual data modalities, which enabled their better integration over time. LEI's design also enabled the identification of features that were consistently important across time for the effective prediction of dementia-related diagnoses. Overall, our work demonstrates the potential of LEI for sequential classification from longitudinal multimodal data., Comment: 11 pages, submitted to ICLR 2025

In a standard navigation pipeline, a robot localizes at every time step to lower navigational errors. However, in some scenarios, a robot needs to selectively localize when it is expensive to obtain observations. For example, an underwater robot surfacing to localize too often hinders it from searching for critical items underwater, such as black boxes from crashed aircraft. On the other hand, if the robot never localizes, poor state estimates cause failure to find the items due to inadvertently leaving the search area or entering hazardous, restricted areas. Motivated by these scenarios, we investigate approaches to help a robot determine "when to localize?" We formulate this as a bi-criteria optimization problem: minimize the number of localization actions while ensuring the probability of failure (due to collision or not reaching a desired goal) remains bounded. In recent work, we showed how to formulate this active localization problem as a constrained Partially Observable Markov Decision Process (POMDP), which was solved using an online POMDP solver. However, this approach is too slow and requires full knowledge of the robot transition and observation models. In this paper, we present RiskRL, a constrained Reinforcement Learning (RL) framework that overcomes these limitations. RiskRL uses particle filtering and recurrent Soft Actor-Critic network to learn a policy that minimizes the number of localizations while ensuring the probability of failure constraint is met. Our numerical experiments show that RiskRL learns a robust policy that outperforms the baseline by at least 13% while also generalizing to unseen environments.

GPT-4o is an autoregressive omni model that accepts as input any combination of text, audio, image, and video, and generates any combination of text, audio, and image outputs. It's trained end-to-end across text, vision, and audio, meaning all inputs and outputs are processed by the same neural network. GPT-4o can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50\% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models. In line with our commitment to building AI safely and consistent with our voluntary commitments to the White House, we are sharing the GPT-4o System Card, which includes our Preparedness Framework evaluations. In this System Card, we provide a detailed look at GPT-4o's capabilities, limitations, and safety evaluations across multiple categories, focusing on speech-to-speech while also evaluating text and image capabilities, and measures we've implemented to ensure the model is safe and aligned. We also include third-party assessments on dangerous capabilities, as well as discussion of potential societal impacts of GPT-4o's text and vision capabilities.

Large Language Models (LLMs) have exhibited strong mathematical reasoning prowess, tackling tasks ranging from basic arithmetic to advanced competition-level problems. However, frequently occurring subtle yet critical errors, such as miscalculations or incorrect substitutions, limit the LLMs' full potential. Existing studies to improve mathematical ability typically involve applying preference learning to step-wise solution pairs. Although these methods leverage samples of varying granularity to mitigate reasoning errors, they overlook critical subtle errors. In this work, we propose a novel preference learning framework called eRror-Injected Self-Editing (RISE), which injects predefined subtle errors into pivotal tokens in reasoning or computation steps to construct hard pairs for error mitigation. In detail, RISE uses the LLM itself to edit a small number of tokens in the solution, injecting designed subtle errors. Then, pairs composed of self-edited solutions and their corresponding correct ones, along with pairs of correct and incorrect solutions obtained through sampling, are used together for subtle error-aware DPO training. Compared with other preference learning methods, RISE further refines the training objective without requiring fine-grained sampling or preference annotation. Extensive experiments validate the effectiveness of RISE, with preference learning on Qwen2-7B-Instruct yielding notable improvements of 3.0% on GSM8K and 7.9% on MATH with only 4.5K training samples. Moreover, the effect of error mitigation extends from mathematical reasoning to logical reasoning and code generation.

Fine-tuning and in-context learning (ICL) are two prevalent methods in imbuing large language models with task-specific knowledge. It is commonly believed that fine-tuning can surpass ICL given sufficient training samples as it allows the model to adjust its internal parameters based on the data. However, this paper presents a counterintuitive finding: For tasks with implicit patterns, ICL captures these patterns significantly better than fine-tuning. We developed several datasets featuring implicit patterns, such as sequences determining answers through parity or identifying reducible terms in calculations. We then evaluated the models' understanding of these patterns under both fine-tuning and ICL across models ranging from 0.5B to 7B parameters. The results indicate that models employing ICL can quickly grasp deep patterns and significantly improve accuracy. In contrast, fine-tuning, despite utilizing thousands of times more training samples than ICL, achieved only limited improvements. We also proposed circuit shift theory from a mechanistic interpretability's view to explain why ICL wins., Comment: EMNLP'24 Findings

Language models are exhibiting increasing capability in knowledge utilization and reasoning. However, when applied as agents in embodied environments, they often suffer from misalignment between their intrinsic knowledge and environmental knowledge, leading to infeasible actions. Traditional environment alignment methods, such as supervised learning on expert trajectories and reinforcement learning, encounter limitations in covering environmental knowledge and achieving efficient convergence, respectively. Inspired by human learning, we propose Exploration-based Error Correction Learning (E2CL), a novel framework that leverages exploration-induced errors and environmental feedback to enhance environment alignment for embodied agents. E2CL incorporates teacher-guided and teacher-free explorations to gather environmental feedback and correct erroneous actions. The agent learns to provide feedback and self-correct, thereby enhancing its adaptability to target environments. Extensive experiments in the VirtualHome environment demonstrate that E2CL-trained agents outperform those trained by baseline methods and exhibit superior self-correction capabilities., Comment: Accepted by EMNLP 2024

Transcribing the speech of multiple overlapping speakers typically requires separating the audio into multiple streams and recognizing each one independently. More recent work jointly separates and transcribes, but requires a separate decoding component for each speaker. We propose the TOGGL model to simultaneously transcribe the speech of multiple speakers. The TOGGL model uses special output tokens to attribute the speech to each speaker with only a single decoder. Our approach generalizes beyond two speakers, even when trained only on two-speaker data. We demonstrate superior performance compared to competing approaches on a conversational speech dataset. Our approach also improves performance on single-speaker audio., Comment: 5 pages

We study an element agglomeration coarsening strategy that requires data redistribution at coarse levels when the number of coarse elements becomes smaller than the used computational units (cores). The overall procedure generates coarse elements (general unstructured unions of fine grid elements) within the framework of element-based algebraic multigrid methods (or AMGe) studied previously. The AMGe generated coarse spaces have the ability to exhibit approximation properties of the same order as the fine-level ones since by construction they contain the piecewise polynomials of the same order as the fine level ones. These approximation properties are key for the successful use of AMGe in multilevel solvers for nonlinear partial differential equations as well as for multilevel Monte Carlo (MLMC) simulations. The ability to coarsen without being constrained by the number of available cores, as described in the present paper, allows to improve the scalability of these solvers as well as in the overall MLMC method. The paper illustrates this latter fact with detailed scalability study of MLMC simulations applied to model Darcy equations with a stochastic log-normal permeability field.

Interrogation of biological pathways is an integral part of omics data analysis. Large language models (LLMs) enable the generation of custom pathways and gene sets tailored to specific scientific questions. These targeted sets are significantly smaller than traditional pathway enrichment analysis libraries, reducing multiple hypothesis testing and potentially enhancing statistical power. Lomics (Large Language Models for Omics Studies) v1.0 is a python-based bioinformatics toolkit that streamlines the generation of pathways and gene sets for transcriptomic analysis. It operates in three steps: 1) deriving relevant pathways based on the researcher's scientific question, 2) generating valid gene sets for each pathway, and 3) outputting the results as .GMX files. Lomics also provides explanations for pathway selections. Consistency and accuracy are ensured through iterative processes, JSON format validation, and HUGO Gene Nomenclature Committee (HGNC) gene symbol verification. Lomics serves as a foundation for integrating LLMs into omics research, potentially improving the specificity and efficiency of pathway analysis.

Previous research has demonstrated the potential of AI agents to act as companions that can provide constant emotional support for humans. In this paper, we emphasize the necessity of autonomous adaptation in personal AI companionship, an underexplored yet promising direction. Such adaptability is crucial as it can facilitate more tailored interactions with users and allow the agent to evolve in response to users' changing needs. However, imbuing agents with autonomous adaptability presents unique challenges, including identifying optimal adaptations to meet users' expectations and ensuring a smooth transition during the adaptation process. To address them, we devise a hierarchical framework, AutoPal, that enables controllable and authentic adjustments to the agent's persona based on user interactions. A personamatching dataset is constructed to facilitate the learning of optimal persona adaptations. Extensive experiments demonstrate the effectiveness of AutoPal and highlight the importance of autonomous adaptability in AI companionship.

The existing safety alignment of Large Language Models (LLMs) is found fragile and could be easily attacked through different strategies, such as through fine-tuning on a few harmful examples or manipulating the prefix of the generation results. However, the attack mechanisms of these strategies are still underexplored. In this paper, we ask the following question: \textit{while these approaches can all significantly compromise safety, do their attack mechanisms exhibit strong similarities?} To answer this question, we break down the safeguarding process of an LLM when encountered with harmful instructions into three stages: (1) recognizing harmful instructions, (2) generating an initial refusing tone, and (3) completing the refusal response. Accordingly, we investigate whether and how different attack strategies could influence each stage of this safeguarding process. We utilize techniques such as logit lens and activation patching to identify model components that drive specific behavior, and we apply cross-model probing to examine representation shifts after an attack. In particular, we analyze the two most representative types of attack approaches: Explicit Harmful Attack (EHA) and Identity-Shifting Attack (ISA). Surprisingly, we find that their attack mechanisms diverge dramatically. Unlike ISA, EHA tends to aggressively target the harmful recognition stage. While both EHA and ISA disrupt the latter two stages, the extent and mechanisms of their attacks differ significantly. Our findings underscore the importance of understanding LLMs' internal safeguarding process and suggest that diverse defense mechanisms are required to effectively cope with various types of attacks., Comment: work in progress

Multi-label image classification datasets are often partially labeled where many labels are missing, posing a significant challenge to training accurate deep classifiers. However, the powerful Mixup sample-mixing data augmentation cannot be well utilized to address this challenge, as it cannot perform linear interpolation on the unknown labels to construct augmented samples. In this paper, we propose LogicMix, a Mixup variant designed for such partially labeled datasets. LogicMix mixes the sample labels by logical OR so that the unknown labels can be correctly mixed by utilizing OR's logical equivalences, including the domination and identity laws. Unlike Mixup, which mixes exactly two samples, LogicMix can mix multiple ($\geq2$) partially labeled samples, constructing visually more confused augmented samples to regularize training. LogicMix is more general and effective than other compared Mixup variants in the experiments on various partially labeled dataset scenarios. Moreover, it is plug-and-play and only requires minimal computation, hence it can be easily inserted into existing frameworks to collaborate with other methods to improve model performance with a negligible impact on training time, as demonstrated through extensive experiments. In particular, through the collaboration of LogicMix, RandAugment, Curriculum Labeling, and Category-wise Fine-Tuning, we attain state-of-the-art performance on MS-COCO, VG-200, and Pascal VOC 2007 benchmarking datasets. The remarkable generality, effectiveness, collaboration, and simplicity suggest that LogicMix promises to be a popular and vital data augmentation method.

Change-point detection and estimation procedures have been widely developed in the literature. However, commonly used approaches in change-point analysis have mainly been focusing on detecting change-points within an entire time series (off-line methods), or quickest detection of change-points in sequentially observed data (on-line methods). Both classes of methods are concerned with change-points that have already occurred. The arguably more important question of when future change-points may occur, remains largely unexplored. In this paper, we develop a novel statistical model that describes the mechanism of change-point occurrence. Specifically, the model assumes a latent process in the form of a random walk driven by non-negative innovations, and an observed process which behaves differently when the latent process belongs to different regimes. By construction, an occurrence of a change-point is equivalent to hitting a regime threshold by the latent process. Therefore, by predicting when the latent process will hit the next regime threshold, future change-points can be forecasted. The probabilistic properties of the model such as stationarity and ergodicity are established. A composite likelihood-based approach is developed for parameter estimation and model selection. Moreover, we construct the predictor and prediction interval for future change points based on the estimated model., Comment: 37 pages, 4 figures

Accurate classification of laryngeal vascular as benign or malignant is crucial for early detection of laryngeal cancer. However, organizations with limited access to laryngeal vascular images face challenges due to the lack of large and homogeneous public datasets for effective learning. Distinguished from the most familiar works, which directly transfer the ImageNet pre-trained models to the target domain for fine-tuning, this work pioneers exploring two-step heterogeneous transfer learning (THTL) for laryngeal lesion classification with nine deep-learning models, utilizing the diabetic retinopathy color fundus images, semantically non-identical yet vascular images, as the intermediate domain. Attention visualization technique, Layer Class Activate Map (LayerCAM), reveals a novel finding that yet the intermediate and the target domain both reflect vascular structure to a certain extent, the prevalent radial vascular pattern in the intermediate domain prevents learning the features of twisted and tangled vessels that distinguish the malignant class in the target domain, summarizes a vital rule for laryngeal lesion classification using THTL. To address this, we introduce an enhanced fine-tuning strategy in THTL called Step-Wise Fine-Tuning (SWFT) and apply it to the ResNet models. SWFT progressively refines model performance by accumulating fine-tuning layers from back to front, guided by the visualization results of LayerCAM. Comparison with the original THTL approach shows significant improvements. For ResNet18, the accuracy and malignant recall increases by 26.1% and 79.8%, respectively, while for ResNet50, these indicators improve by 20.4% and 62.2%, respectively.

The research field of Artificial Life studies how life-like phenomena such as autopoiesis, agency, or self-regulation can self-organize in computer simulations. In cellular automata (CA), a key open-question has been whether it it is possible to find environment rules that self-organize robust "individuals" from an initial state with no prior existence of things like "bodies", "brain", "perception" or "action". In this paper, we leverage recent advances in machine learning, combining algorithms for diversity search, curriculum learning and gradient descent, to automate the search of such "individuals", i.e. localized structures that move around with the ability to react in a coherent manner to external obstacles and maintain their integrity, hence primitive forms of sensorimotor agency. We show that this approach enables to find systematically environmental conditions in CA leading to self-organization of such basic forms of agency. Through multiple experiments, we show that the discovered agents have surprisingly robust capabilities to move, maintain their body integrity and navigate among various obstacles. They also show strong generalization abilities, with robustness to changes of scale, random updates or perturbations from the environment not seen during training. We discuss how this approach opens new perspectives in AI and synthetic bioengineering.

Tuning language models for dialogue generation has been a prevalent paradigm for building capable dialogue agents. Yet, traditional tuning narrowly views dialogue generation as resembling other language generation tasks, ignoring the role disparities between two speakers and the multi-round interactive process that dialogues ought to be. Such a manner often leads to unsatisfactory chat consistency for the built agent. In this work, we emphasize the interactive, communicative nature of dialogue and argue that it is more feasible to model the speaker roles of agent and user separately, enabling the agent to adhere to its role consistently. With this in mind, we propose an efficient Multi-round Interactive Dialogue Tuning (Midi-Tuning) framework. It models the agent and user individually with two adapters built upon large language models. The adapters make use of respective utterances round by round in alternating order and they are tuned via a round-level memory caching mechanism. Extensive experiments demonstrate that, our framework performs superior to traditional fine-tuning and harbors the tremendous potential for improving dialogue consistency., Comment: Accepted by ACL 2024

Black hole dynamics suggests that dark matter would re-distribute near a supermassive black hole to form a density spike. However, no direct evidence of dark matter density spike around a supermassive black hole has been identified. In this letter, we present the first robust evidence showing a dark matter density spike around a supermassive black hole. We revisit the data of the well-known supermassive black hole binary OJ 287 and show that the inclusion of the dynamical friction due to a dark matter density spike around the supermassive black hole can satisfactorily account for the observed orbital decay rate. The derived spike index $\gamma_{\rm sp}=2.351^{+0.032}_{-0.045}$ gives an excellent agreement with the value $\gamma_{\rm sp}=2.333$ predicted by the benchmark model assuming an adiabatically growing supermassive black hole. This provides a strong verification of the canonical theory suggested two decades ago modeling the gravitational interaction between collisionless dark matter and supermassive black holes., Comment: Accepted for publication in ApJL

Large-scale image datasets are often partially labeled, where only a few categories' labels are known for each image. Assigning pseudo-labels to unknown labels to gain additional training signals has become prevalent for training deep classification models. However, some pseudo-labels are inevitably incorrect, leading to a notable decline in the model classification performance. In this paper, we propose a novel method called Category-wise Fine-Tuning (CFT), aiming to reduce model inaccuracies caused by the wrong pseudo-labels. In particular, CFT employs known labels without pseudo-labels to fine-tune the logistic regressions of trained models individually to calibrate each category's model predictions. Genetic Algorithm, seldom used for training deep models, is also utilized in CFT to maximize the classification performance directly. CFT is applied to well-trained models, unlike most existing methods that train models from scratch. Hence, CFT is general and compatible with models trained with different methods and schemes, as demonstrated through extensive experiments. CFT requires only a few seconds for each category for calibration with consumer-grade GPUs. We achieve state-of-the-art results on three benchmarking datasets, including the CheXpert chest X-ray competition dataset (ensemble mAUC 93.33%, single model 91.82%), partially labeled MS-COCO (average mAP 83.69%), and Open Image V3 (mAP 85.31%), outperforming the previous bests by 0.28%, 2.21%, 2.50%, and 0.91%, respectively. The single model on CheXpert has been officially evaluated by the competition server, endorsing the correctness of the result. The outstanding results and generalizability indicate that CFT could be substantial and prevalent for classification model development. Code is available at: https://github.com/maxium0526/category-wise-fine-tuning.

In a binary, when the orbital plane of the companion star is almost edge-on along the line-of-sight direction, this would produce an observable self-gravitational lensing effect, which would slightly increase the overall optical intensity of the binary. However, the probability of getting one observable self-lensing binary (SLB) is very low. There are only five observed SLBs so far and all of them are eclipsing binaries. In this article, we theoretically show that the neutron star-white dwarf (NS-WD) binary PSR J1910-5959A could be an observable non-eclipsing SLB. It might be the first binary showing both periodic optical amplification and Shapiro time delay of radio signals, which is useful to verify our understanding about gravitational lensing in relativistic binaries. Moreover, we show that the observed peak amplification limit of the PSR J1910-5959A can help constrain the radius of the WD, which is a crucial parameter to examine the mass-radius and temperature-radius relationship for helium WD., Comment: Accepted for publication in Physical Review D

Radio data can give stringent constraints for annihilating dark matter. In general, radio observations can detect very accurate radio flux density with high resolution and different frequencies for nearby galaxies. We are able to obtain the radio flux density as a function of distance from the galactic center and frequencies $S(r,\nu)$. In this article, we demonstrate a comprehensive radio analysis of the M33 galaxy, combining the radio flux density profile $S(r)$ and the frequency spectrum $S(\nu)$ to get the constraints of dark matter annihilation parameters. By analyzing the archival radio data obtained from the Effelsberg telescope, we show that the dark matter annihilation contributing to the radio flux density might be insignificant in the disk region of the M33 galaxy. Moreover, by including the baryonic radio contribution, we constrain the $2\sigma$ conservative upper limits of the annihilation cross section, which can be complementary to the existing constraints based on neutrino, cosmic-ray, and gamma-ray observations. Our results indicate that analyzing the galactic multi-frequency radio flux profiles can give useful and authentic constraints on dark matter for the leptophilic annihilation channels., Comment: Accepted publication in ApJ

In this paper, we introduce eipy--an open-source Python package for developing effective, multi-modal heterogeneous ensembles for classification. eipy simultaneously provides both a rigorous, and user-friendly framework for comparing and selecting the best-performing multi-modal data integration and predictive modeling methods by systematically evaluating their performance using nested cross-validation. The package is designed to leverage scikit-learn-like estimators as components to build multi-modal predictive models. An up-to-date user guide, including API reference and tutorials, for eipy is maintained at https://eipy.readthedocs.io . The main repository for this project can be found on GitHub at https://github.com/GauravPandeyLab/eipy .