The electric power sector is a leading source of air pollutant emissions, impacting the public health of nearly every community. Although regulatory measures have reduced air pollutants, fossil fuels remain a significant component of the energy supply, highlighting the need for more advanced demand-side approaches to reduce the public health impacts. To enable health-informed demand-side management, we introduce HealthPredictor, a domain-specific AI model that provides an end-to-end pipeline linking electricity use to public health outcomes. The model comprises three components: a fuel mix predictor that estimates the contribution of different generation sources, an air quality converter that models pollutant emissions and atmospheric dispersion, and a health impact assessor that translates resulting pollutant changes into monetized health damages. Across multiple regions in the United States, our health-driven optimization framework yields substantially lower prediction errors in terms of public health impacts than fuel mix-driven baselines. A case study on electric vehicle charging schedules illustrates the public health gains enabled by our method and the actionable guidance it can offer for health-informed energy management. Overall, this work shows how AI models can be explicitly designed to enable health-informed energy management for advancing public health and broader societal well-being. Our datasets and code are released at: https://github.com/Ren-Research/Health-Impact-Predictor.
Time series foundation models (TSFMs) based on autoregressive (AR) next-token prediction entangle two factors: variable-context conditioning and prediction-unit factorization with autoregressive rollout. Strong AR performance therefore does not reveal whether generalization stems from the flexible training distribution, the local prediction target, the left-to-right inference process, or their combination. We propose FlexCast, a flexible context-target learning framework that disentangles these factors. FlexCast reformulates forecasting as variable-length masked prediction over time series patches: during pretraining, it dynamically samples visible context and masked target patches, exposing the model to diverse context-target configurations without sequential decoding. At inference time, the same pretrained model supports one-step forecasting, confidence-based iterative refinement, and next-token-style generation through a unified mask-and-predict interface, enabling controlled comparison of inference schedules under identical architecture, data, and objective. Evaluated on BLAST and multiple zero-shot benchmarks, FlexCast achieves accuracy competitive with or better than strong TSFM baselines while substantially reducing inference latency. Ablations confirm that context diversity and contiguous target masking drive performance, while left-to-right decoding does not yield consistent gains once flexible context-target relationships have been learned, suggesting that the key benefit of AR-style TSFMs lies in flexible training rather than sequential inference. The code is available at https://github.com/BenchCouncil/FlexCast.
기존 비전-언어 내비게이션(VLN) 방법들은 다단계 행동 계획을 생성하기 위해 VLM을 웨이포인트 디코더와 결합하는 경우가 많지만, 장기 지평 내비게이션에서 의미론적 진행을 추적하고 실행 실패를 진단하며 오류 누적으로부터 복구하기 위한 명시적인 폐루프(closed-loop) 메커니즘이 대체로 부재하다. 이 공백을 메우기 위해, 우리는 의도 에이전트와 실행 에이전트 간의 양방향 상호작용을 통해 의사결정을 조직화하는 에이전트형 VLN 프레임워크인 ReflectVLN을 제안한다. 의도 에이전트는 하위 작업 분해와 반성(reflection)을 수행하며, 수정적 계획으로 실행 가능한 하위 작업 설명을 생성한다. 이러한 설명을 조건으로, 실행 에이전트는 현재 관측 하에서 이를 단기 지평 행동으로 그라운딩(grounding)하면서 하위 목표 진행을 모니터링하고 경로 이탈 행동을 감지한다. 핵심적으로, ReflectVLN은 폐루프 양방향 통신을 가능하게 한다. 실행 에이전트는 필요 시 반성과 하위 작업 갱신을 촉발하는 진행 및 편차 신호를 방출하고, 의도 에이전트는 복구를 위해 후속 행동을 재조건화하는 구조화된 지침을 반환한다. 해석 가능한 중간 근거를 갖춘 시간적으로 일관된 의사결정을 장려하기 위해, 우리는 행동 생성을 위한 경로 조건 이중 질의 학습 방식인 Action Chain-of-Thought(Action-CoT)를 도입한다. 표준 VLN 벤치마크에 대한 실험 결과, ReflectVLN은 제한된 데이터 예산 하에서 성공률과 경로 효율을 개선하며, 유리한 학습 비용과 추론 시 더 적은 고수준 의도 호출을 달성하면서, 분석 및 협업을 위한 해석 가능한 중간 의사결정을 제공한다. 코드는 다음에서 이용 가능하다: https://github.com/AIprogrammer/ReflectVLN
심층 검색(deep search)은 개방형 도메인의 복잡한 작업을 해결하기 위한 대규모 언어 모델(LLM)의 핵심 역량이 되었습니다. 그러나 기존 접근 방식은 일반적으로 궤적 생성과 추론(inference) 모두에서 선형적 순차 추론에 의존하므로, 긴 지평(long-horizon)의 다중 홉(multi-hop) 검색 전반에 걸쳐 중간 상태와 제약 조건을 일관되게 보존하기 어렵습니다. 결과적으로 이들은 문맥 망각, 검색 표류(search drift), 비효율적인 탐험(exploration) 문제를 자주 겪습니다. 이러한 한계를 해결하기 위해, 우리는 추론 과정을 고정 토폴로지 쿼리 그래프 상의 상태 진화로 구조화하는 심층 검색용 추론 프레임워크인 G-ReAct를 제안합니다. 진화하는 그래프 상태는 검색 진행 상황을 명시적으로 추적하고 후속 결정을 유도하여, 텍스트 이력에 의해 주도되던 탐색적 검색을 명시적 제약 조건 하의 그래프 유도 추론으로 변환합니다. G-ReAct는 학습(training)과 추론을 모두 지원합니다. 즉, 지도 미세조정(supervised fine-tuning)을 위한 고품질 심층 검색 궤적을 생성하며, 추가 미세조정(fine-tuning) 없이도 추론 시점 검색을 위한 구조화된 지침을 제공합니다. 실험을 통해 미세조정을 위한 단 1.9K개의 생성된 궤적만으로도 Qwen3-30B-A3B-Thinking-2507이 BrowseComp-ZH에서 52.6%, XBench에서 79.0%의 정확도를 달성하여 강화학습(RL) 기반 방식을 포함해 실질적으로 더 큰 데이터셋으로 학습된 유사한 오픈소스 방식들보다 뛰어난 성능을 보임을 입증합니다. 또한, 추론 시점에 적용될 때 G-ReAct는 심층 검색 작업에서 기존의 강력한 LLM 성능을 일관되게 향상시킵니다. 모든 코드와 모델(model) 가중치는 공개될 예정입니다.
뉴로모픽 센서인 이벤트 카메라는 높은 시간 해상도와 넓은 동적 범위(high dynamic range)라는 장점을 지니며, 밝기 변화를 희소한 이벤트 스트림 형태로 비동기식으로 기록합니다. 이벤트로부터 강도(intensity) 이미지를 재구성하는 작업은 절대 밝기에 대한 본질적인 모호성으로 인해 매우 불량 설정된(ill-posed) 문제입니다. 초기 방법들은 대체로 종단간(end-to-end) 회귀 패러다임을 따라 이벤트를 결정론적 방식으로 강도 프레임에 직접 매핑했습니다. 이러한 접근법은 어느 정도 효과적이었으나 지각적으로 열등한 결과를 내는 경우가 많았으며, 모델 용량 및 학습(training) 데이터의 확장(scaling)에 어려움을 겪었습니다. 본 논문에서는 대리 학습(surrogate training) 프레임워크를 따라 고품질 비디오를 생성하는 이벤트 기반 확산 모델(diffusion model)인 EvDiff를 제안합니다. 고프레임률 비디오 생성 시 발생하는 높은 연산 비용을 줄이기 위해, 시간적 일관성을 유지하는 EvEncoder를 갖추고 단 한 번의 전방 확산 단계만을 수행하는 이벤트 기반 확산 모델을 설계했습니다. 나아가 당사의 새로운 대리 학습 프레임워크는 쌍을 이루는 이벤트-이미지 데이터셋(dataset)에 대한 의존성을 제거하여 모델(model)이 더 큰 용량을 위해 대규모 이미지 데이터셋을 활용할 수 있도록 합니다. 제안된 EvDiff는 단색 이벤트 스트림만으로도 고품질 컬러 비디오를 생성할 수 있습니다. 실제 데이터셋에 대한 실험을 통해 당사의 방법이 충실도(fidelity)와 사실감(realism) 사이의 최적 지점을 달성하며, 픽셀 수준 및 지각적 지표 모두에서 기존 접근법을 능가함을 입증했습니다. 코드는 공개될 예정입니다.
4족 보행(quadrupedal locomotion)은 명령 추적(command tracking), 안정성, 에너지 효율성과 같은 상충하는 목표들 사이의 균형을 필요로 하지만, 기존의 강화학습(RL)은 이러한 우선순위를 학습 시점에 고정된 스칼라 보상으로 하드코딩한다. 우리는 이러한 트레이드오프를 단일 보행 정책(policy)에 대한 명시적인 런타임 입력으로 만드는 시맨틱 다목적 접근법인 PROMO(Preference-Conditioned Multi-Objective Reinforcement Learning, 선호도 조건부 다목적 강화학습)를 제시한다. PROMO는 체화(embodiment) 특유의 보행 사전 지식(locomotion prior)은 고정한 채, 배포 시점의 선호도(preference)에 정책을 조건화함으로써, 운용자의 의도를 실행 가능한 보행 패턴(gait) 생성에 필요한 보상 형성(reward shaping) 항과 분리한다. 고정 목표 제어기, 다목적 베이스라인, 독립적으로 학습된 전문가 정책(specialist)과 비교했을 때, PROMO는 단일 배포 가능 정책만으로 목표별 특화(specialization)와 강건성(robustness)을 달성한다. 시뮬레이션에서 샘플링한 100개의 선호도 중 67개의 행동이 엄밀한 파레토 지배(Pareto dominance) 기준으로 비지배(non-dominated) 상태였으며, 평균 선호도-목표 상관관계(correlation)는 0.843으로, 넓은 파레토 커버리지(coverage)와 예측 가능한 선호도 반응을 보여준다. 동일한 정책은 제로샷(zero-shot)으로 Unitree Go2에 전이(transfer)되며, 선호도 변경만으로 균형 잡힌 선호도 대비 비에너지(specific energy)는 최대 30.4%, 위치 오차는 38.7%, 최대 자세 편차(peak body-attitude deviation)는 59.0%까지 감소시킨다. 이러한 결과는 선호도 조건부 다목적 강화학습이 적응형 다리형 보행(legged locomotion)을 위한 실용적인 런타임 인터페이스로 자리매김할 수 있음을 보여주며, 그 역할을 오프라인 파레토 집합(Pareto-set) 구성을 넘어 확장한다. 오픈소스 코드와 영상은 https://amrmousa.com/promo/ 에서 확인할 수 있다.
The success of generative modeling in continuous domains has led to a surge of interest in generating discrete data such as molecules, source code, and graphs. However, construction histories for these discrete objects are typically not unique and so generative models must reason about intractably large spaces in order to learn. Additionally, structured discrete domains are often characterized by strict constraints on what constitutes a valid object and generative models must respect these requirements in order to produce useful novel samples. Here, we present a generative model for discrete objects employing a Markov chain where transitions are restricted to a set of local operations that preserve validity. Building off of generative interpretations of denoising autoencoders, the Markov chain alternates between producing 1) a sequence of corrupted objects that are valid but not from the data distribution, and 2) a learned reconstruction distribution that attempts to fix the corruptions while also preserving validity. This approach constrains the generative model to only produce valid objects, requires the learner to only discover local modifications to the objects, and avoids marginalization over an unknown and potentially large space of construction histories. We evaluate the proposed approach on two highly structured discrete domains, molecules and Laman graphs, and find that it compares favorably to alternative methods at capturing distributional statistics for a host of semantically relevant metrics.
The synthesis of human grasping has numerous applications including AR/VR, video games and robotics. While methods have been proposed to generate realistic hand-object interaction for object grasping and manipulation, these typically only consider interacting hand alone. Our goal is to synthesize whole-body grasping motions. Starting from an arbitrary initial pose, we aim to generate diverse and natural whole-body human motions to approach and grasp a target object in 3D space. This task is challenging as it requires modeling both whole-body dynamics and dexterous finger movements. To this end, we propose SAGA (StochAstic whole-body Grasping with contAct), a framework which consists of two key components: (a) Static whole-body grasping pose generation. Specifically, we propose a multi-task generative model, to jointly learn static whole-body grasping poses and human-object contacts. (b) Grasping motion infilling. Given an initial pose and the generated whole-body grasping pose as the start and end of the motion respectively, we design a novel contact-aware generative motion infilling module to generate a diverse set of grasp-oriented motions. We demonstrate the effectiveness of our method, which is a novel generative framework to synthesize realistic and expressive whole-body motions that approach and grasp randomly placed unseen objects. Code and models are available at https://jiahaoplus.github.io/SAGA/saga.html.
Speech representations learned from Self-supervised learning (SSL) models can benefit various speech processing tasks. However, utilizing SSL representations usually requires fine-tuning the pre-trained models or designing task-specific downstream models and loss functions, causing much memory usage and human labor. Recently, prompting in Natural Language Processing (NLP) has been found to be an efficient technique to leverage pre-trained language models (LMs). Specifically, prompt tuning optimizes a limited number of task-specific parameters with a fixed pre-trained model; as a result, only a small set of parameters is needed to be stored for each task. Prompt tuning improves computation and memory efficiency by leveraging the pre-trained LM's prediction ability. Nevertheless, such a paradigm is little studied in the speech community. We report in this paper the first exploration of the prompt tuning paradigm for speech processing tasks based on Generative Spoken Language Model (GSLM). Experiment results show that the prompt tuning technique achieves competitive performance in speech classification tasks with fewer trainable parameters than fine-tuning specialized downstream models. We further study the technique in challenging sequence generation tasks. Prompt tuning also demonstrates its potential, while the limitation and possible research directions are discussed in this paper. The source code is available on https://github.com/ga642381/SpeechPrompt.
Generating realistic 3D worlds occupied by moving humans has many applications in games, architecture, and synthetic data creation. But generating such scenes is expensive and labor intensive. Recent work generates human poses and motions given a 3D scene. Here, we take the opposite approach and generate 3D indoor scenes given 3D human motion. Such motions can come from archival motion capture or from IMU sensors worn on the body, effectively turning human movement in a "scanner" of the 3D world. Intuitively, human movement indicates the free-space in a room and human contact indicates surfaces or objects that support activities such as sitting, lying or touching. We propose MIME (Mining Interaction and Movement to infer 3D Environments), which is a generative model of indoor scenes that produces furniture layouts that are consistent with the human movement. MIME uses an auto-regressive transformer architecture that takes the already generated objects in the scene as well as the human motion as input, and outputs the next plausible object. To train MIME, we build a dataset by populating the 3D FRONT scene dataset with 3D humans. Our experiments show that MIME produces more diverse and plausible 3D scenes than a recent generative scene method that does not know about human movement. Code and data will be available for research at https://mime.is.tue.mpg.de.
Synthesizing high-fidelity complex images from text is challenging. Based on large pretraining, the autoregressive and diffusion models can synthesize photo-realistic images. Although these large models have shown notable progress, there remain three flaws. 1) These models require tremendous training data and parameters to achieve good performance. 2) The multi-step generation design slows the image synthesis process heavily. 3) The synthesized visual features are difficult to control and require delicately designed prompts. To enable high-quality, efficient, fast, and controllable text-to-image synthesis, we propose Generative Adversarial CLIPs, namely GALIP. GALIP leverages the powerful pretrained CLIP model both in the discriminator and generator. Specifically, we propose a CLIP-based discriminator. The complex scene understanding ability of CLIP enables the discriminator to accurately assess the image quality. Furthermore, we propose a CLIP-empowered generator that induces the visual concepts from CLIP through bridge features and prompts. The CLIP-integrated generator and discriminator boost training efficiency, and as a result, our model only requires about 3% training data and 6% learnable parameters, achieving comparable results to large pretrained autoregressive and diffusion models. Moreover, our model achieves 120 times faster synthesis speed and inherits the smooth latent space from GAN. The extensive experimental results demonstrate the excellent performance of our GALIP. Code is available at https://github.com/tobran/GALIP.
In this paper, we improve the challenging monocular 3D object detection problem with a general semi-supervised framework. Specifically, having observed that the bottleneck of this task lies in lacking reliable and informative samples to train the detector, we introduce a novel, simple, yet effective `Augment and Criticize' framework that explores abundant informative samples from unlabeled data for learning more robust detection models. In the `Augment' stage, we present the Augmentation-based Prediction aGgregation (APG), which aggregates detections from various automatically learned augmented views to improve the robustness of pseudo label generation. Since not all pseudo labels from APG are beneficially informative, the subsequent `Criticize' phase is presented. In particular, we introduce the Critical Retraining Strategy (CRS) that, unlike simply filtering pseudo labels using a fixed threshold (e.g., classification score) as in 2D semi-supervised tasks, leverages a learnable network to evaluate the contribution of unlabeled images at different training timestamps. This way, the noisy samples prohibitive to model evolution could be effectively suppressed. To validate our framework, we apply it to MonoDLE and MonoFlex. The two new detectors, dubbed 3DSeMo_DLE and 3DSeMo_FLEX, achieve state-of-the-art results with remarkable improvements for over 3.5% AP_3D/BEV (Easy) on KITTI, showing its effectiveness and generality. Code and models will be released.
The art of mathematical reasoning stands as a fundamental pillar of intellectual progress and is a central catalyst in cultivating human ingenuity. Researchers have recently published a plethora of works centered around the task of solving Math Word Problems (MWP) $-$ a crucial stride towards general AI. These existing models are susceptible to dependency on shallow heuristics and spurious correlations to derive the solution expressions. In order to ameliorate this issue, in this paper, we propose a framework for MWP solvers based on the generation of linguistic variants of the problem text. The approach involves solving each of the variant problems and electing the predicted expression with the majority of the votes. We use DeBERTa (Decoding-enhanced BERT with disentangled attention) as the encoder to leverage its rich textual representations and enhanced mask decoder to construct the solution expressions. Furthermore, we introduce a challenging dataset, $\mathrm{P\small{ARA}\normalsize{MAWPS}}$, consisting of paraphrased, adversarial, and inverse variants of selectively sampled MWPs from the benchmark $\mathrm{M\small{AWPS}}$ dataset. We extensively experiment on this dataset along with other benchmark datasets using some baseline MWP solver models. We show that training on linguistic variants of problem statements and voting on candidate predictions improve the mathematical reasoning and robustness of the model. We make our code and data publicly available.
Robust optimization is an established framework for modeling optimization problems with uncertain parameters. While static robust optimization is often criticized for being too conservative, two-stage (or adjustable) robust optimization (2RO) provides a less conservative alternative by allowing some decisions to be made after the uncertain parameters have been revealed. Unfortunately, in the case of integer decision variables, existing solution methods for 2RO typically fail to solve large-scale instances, limiting the applicability of this modeling paradigm to simple cases. We propose Neur2RO, a deep-learning-augmented instantiation of the column-and-constraint-generation (CCG) algorithm, which expands the applicability of the 2RO framework to large-scale instances with integer decisions in both stages. A custom-designed neural network is trained to estimate the optimal value and feasibility of the second-stage problem. The network can be incorporated into CCG, leading to more computationally tractable subproblems in each of its iterations. The resulting algorithm enjoys approximation guarantees which depend on the neural network's prediction error. In our experiments, Neur2RO produces high-quality solutions quickly, outperforming state-of-the-art methods on two-stage knapsack, capital budgeting, and facility location problems. Compared to existing methods, which often run for hours, Neur2RO finds better solutions in a few seconds or minutes. Our code is available at https://github.com/khalil-research/Neur2RO.
Watermark algorithms for large language models (LLMs) have achieved extremely high accuracy in detecting text generated by LLMs. Such algorithms typically involve adding extra watermark logits to the LLM's logits at each generation step. However, prior algorithms face a trade-off between attack robustness and security robustness. This is because the watermark logits for a token are determined by a certain number of preceding tokens; a small number leads to low security robustness, while a large number results in insufficient attack robustness. In this work, we propose a semantic invariant watermarking method for LLMs that provides both attack robustness and security robustness. The watermark logits in our work are determined by the semantics of all preceding tokens. Specifically, we utilize another embedding LLM to generate semantic embeddings for all preceding tokens, and then these semantic embeddings are transformed into the watermark logits through our trained watermark model. Subsequent analyses and experiments demonstrated the attack robustness of our method in semantically invariant settings: synonym substitution and text paraphrasing settings. Finally, we also show that our watermark possesses adequate security robustness. Our code and data are available at \href{https://github.com/THU-BPM/Robust_Watermark}{https://github.com/THU-BPM/Robust\_Watermark}. Additionally, our algorithm could also be accessed through MarkLLM \citep{pan2024markllm} \footnote{https://github.com/THU-BPM/MarkLLM}.
Current RGBT tracking research relies on the complete multi-modal input, but modal information might miss due to some factors such as thermal sensor self-calibration and data transmission error, called modality-missing challenge in this work. To address this challenge, we propose a novel invertible prompt learning approach, which integrates the content-preserving prompts into a well-trained tracking model to adapt to various modality-missing scenarios, for robust RGBT tracking. Given one modality-missing scenario, we propose to utilize the available modality to generate the prompt of the missing modality to adapt to RGBT tracking model. However, the cross-modality gap between available and missing modalities usually causes semantic distortion and information loss in prompt generation. To handle this issue, we design the invertible prompter by incorporating the full reconstruction of the input available modality from the generated prompt. To provide a comprehensive evaluation platform, we construct several high-quality benchmark datasets, in which various modality-missing scenarios are considered to simulate real-world challenges. Extensive experiments on three modality-missing benchmark datasets show that our method achieves significant performance improvements compared with state-of-the-art methods. We have released the code and simulation datasets at: \href{https://github.com/Alexadlu/Modality-missing-RGBT-Tracking.git}{https://github.com/Alexadlu/Modality-missing-RGBT-Tracking.git}.
We introduce Garment3DGen a new method to synthesize 3D garment assets from a base mesh given a single input image as guidance. Our proposed approach allows users to generate 3D textured clothes based on both real and synthetic images, such as those generated by text prompts. The generated assets can be directly draped and simulated on human bodies. We leverage the recent progress of image-to-3D diffusion methods to generate 3D garment geometries. However, since these geometries cannot be utilized directly for downstream tasks, we propose to use them as pseudo ground-truth and set up a mesh deformation optimization procedure that deforms a base template mesh to match the generated 3D target. Carefully designed losses allow the base mesh to freely deform towards the desired target, yet preserve mesh quality and topology such that they can be simulated. Finally, we generate high-fidelity texture maps that are globally and locally consistent and faithfully capture the input guidance, allowing us to render the generated 3D assets. With Garment3DGen users can generate the simulation-ready 3D garment of their choice without the need of artist intervention. We present a plethora of quantitative and qualitative comparisons on various assets and demonstrate that Garment3DGen unlocks key applications ranging from sketch-to-simulated garments or interacting with the garments in VR. Code is publicly available.
The progress in deep learning solutions for disease diagnosis and prognosis based on cardiac magnetic resonance imaging is hindered by highly imbalanced and biased training data. To address this issue, we propose a method to alleviate imbalances inherent in datasets through the generation of synthetic data based on sensitive attributes such as sex, age, body mass index, and health condition. We adopt ControlNet based on a denoising diffusion probabilistic model to condition on text assembled from patient metadata and cardiac geometry derived from segmentation masks using a large-cohort study, specifically, the UK Biobank. We assess our method by evaluating the realism of the generated images using established quantitative metrics. Furthermore, we conduct a downstream classification task aimed at debiasing a classifier by rectifying imbalances within underrepresented groups through synthetically generated samples. Our experiments demonstrate the effectiveness of the proposed approach in mitigating dataset imbalances, such as the scarcity of younger patients or individuals with normal BMI level suffering from heart failure. This work represents a major step towards the adoption of synthetic data for the development of fair and generalizable models for medical classification tasks. Notably, we conduct all our experiments using a single, consumer-level GPU to highlight the feasibility of our approach within resource-constrained environments. Our code is available at https://github.com/faildeny/debiasing-cardiac-mri.
The temperature parameter plays a profound role during training and/or inference with large foundation models (LFMs) such as large language models (LLMs) and CLIP models. Particularly, it adjusts the logits in the softmax function in LLMs, which is crucial for next token generation, and it scales the similarities in the contrastive loss for training CLIP models. A significant question remains: Is it viable to learn a neural network to predict a personalized temperature of any input data for enhancing LFMs"? In this paper, we present a principled framework for learning a small yet generalizable temperature prediction network (TempNet) to improve LFMs. Our solution is composed of a novel learning framework with a robust loss underpinned by constrained distributionally robust optimization (DRO), and a properly designed TempNet with theoretical inspiration. TempNet can be trained together with a large foundation model from scratch or learned separately given a pretrained foundation model. It is not only useful for predicting personalized temperature to promote the training of LFMs but also generalizable and transferable to new tasks. Our experiments on LLMs and CLIP models demonstrate that TempNet greatly improves the performance of existing solutions or models, e.g. Table 1. The code to reproduce the experimental results in this paper can be found at https://github.com/zhqiu/TempNet.
State-of-the-art deep learning entity linking methods rely on extensive human-labelled data, which is costly to acquire. Current datasets are limited in size, leading to inadequate coverage of biomedical concepts and diminished performance when applied to new data. In this work, we propose to automatically generate data to create large-scale training datasets, which allows the exploration of approaches originally developed for the task of extreme multi-label ranking in the biomedical entity linking task. We propose the hybrid X-Linker pipeline that includes different modules to link disease and chemical entity mentions to concepts in the MEDIC and the CTD-Chemical vocabularies, respectively. X-Linker was evaluated on several biomedical datasets: BC5CDR-Disease, BioRED-Disease, NCBI-Disease, BC5CDR-Chemical, BioRED-Chemical, and NLM-Chem, achieving top-1 accuracies of 0.8307, 0.7969, 0.8271, 0.9511, 0.9248, and 0.7895, respectively. X-Linker demonstrated superior performance in three datasets: BC5CDR-Disease, NCBI-Disease, and BioRED-Chemical. In contrast, SapBERT outperformed X-Linker in the remaining three datasets. Both models rely only on the mention string for their operations. The source code of X-Linker and its associated data are publicly available for performing biomedical entity linking without requiring pre-labelled entities with identifiers from specific knowledge organization systems.