This tutorial provides a comprehensive survey of methods for fine-tuning diffusion models to optimize downstream reward functions. While diffusion models are widely known to provide excellent generative modeling capability, practical applications in domains such as biology require generating samples that maximize some desired metric (e.g., translation efficiency in RNA, docking score in molecules, stability in protein). In these cases, the diffusion model can be optimized not only to generate realistic samples but also to explicitly maximize the measure of interest. Such methods are based on concepts from reinforcement learning (RL). We explain the application of various RL algorithms, including PPO, differentiable optimization, reward-weighted MLE, value-weighted sampling, and path consistency learning, tailored specifically for fine-tuning diffusion models. We aim to explore fundamental aspects such as the strengths and limitations of different RL-based fine-tuning algorithms across various scenarios, the benefits of RL-based fine-tuning compared to non-RL-based approaches, and the formal objectives of RL-based fine-tuning (target distributions). Additionally, we aim to examine their connections with related topics such as classifier guidance, Gflownets, flow-based diffusion models, path integral control theory, and sampling from unnormalized distributions such as MCMC. The code of this tutorial is available at https://github.com/masa-ue/RLfinetuning_Diffusion_Bioseq

Parameter-efficient tunings (PETs) have demonstrated impressive performance and promising perspectives in training large models, while they are still confronted with a common problem: the trade-off between learning new content and protecting old knowledge, leading to zero-shot generalization collapse, and cross-modal hallucination. In this paper, we reformulate Adapter, LoRA, Prefix-tuning, and Prompt-tuning from the perspective of gradient projection, and firstly propose a unified framework called Parameter Efficient Gradient Projection (PEGP). We introduce orthogonal gradient projection into different PET paradigms and theoretically demonstrate that the orthogonal condition for the gradient can effectively resist forgetting even for large-scale models. It therefore modifies the gradient towards the direction that has less impact on the old feature space, with less extra memory space and training time. We extensively evaluate our method with different backbones, including ViT and CLIP, on diverse datasets, and experiments comprehensively demonstrate its efficiency in reducing forgetting in class, online class, domain, task, and multi-modality continual settings. The project page is available at https://dmcv-ecnu-pegp.github.io/.

Comprehensive sex education (SE) is crucial in promoting sexual health and responsible behavior among students, particularly in elementary schools. Despite its significance, teaching SE can be challenging due to students' attitudes, shyness, and emotional barriers. Socially assistive robots (SARs) sometimes are perceived as more trustworthy than humans, based on research showing that they are not anticipated as judgmental. Inspired by those evidences, this study aims to assess the success of a SAR as a facilitator for SE lessons for elementary school students. This study conducted two experiments to assess the effectiveness of a SAR in facilitating SE education for elementary school students. We conducted two experiments, a) a group activity in the school classroom where the Nao robot gave a SE lecture, and we evaluated how much information the students acquired from the lecture, and b) an individual activity where the students interacted 1:1 with the robot, and we evaluated their attitudes towards the subject of SE, and if they felt comfortable to ask SE related questions to the robot. Data collected from pre- and post-questionnaires, as well as video annotations, revealed that the SAR significantly improved students' attitudes toward SE. Furthermore, students were more open to asking SE-related questions to the robot than their human teacher. The study emphasized specific SAR characteristics, such as embodiment and non-judgmental behavior, as key factors contributing to their effectiveness in supporting SE education, paving the way for innovative and effective approaches to sexual education in schools.

Combining offline and online reinforcement learning (RL) techniques is indeed crucial for achieving efficient and safe learning where data acquisition is expensive. Existing methods replay offline data directly in the online phase, resulting in a significant challenge of data distribution shift and subsequently causing inefficiency in online fine-tuning. To address this issue, we introduce an innovative approach, \textbf{E}nergy-guided \textbf{DI}ffusion \textbf{S}ampling (EDIS), which utilizes a diffusion model to extract prior knowledge from the offline dataset and employs energy functions to distill this knowledge for enhanced data generation in the online phase. The theoretical analysis demonstrates that EDIS exhibits reduced suboptimality compared to solely utilizing online data or directly reusing offline data. EDIS is a plug-in approach and can be combined with existing methods in offline-to-online RL setting. By implementing EDIS to off-the-shelf methods Cal-QL and IQL, we observe a notable 20% average improvement in empirical performance on MuJoCo, AntMaze, and Adroit environments. Code is available at \url{https://github.com/liuxhym/EDIS}.

Objective B. Messages in the forum are summarized using automatic natural language techniques. These summaries are intended to help students identify the errors they are having and improve their ability to ask for help. When they use the selected recommendations, students will be provided with strategies to learn how to ask better questions iteratively. "Improving question formulation" represents acquiring a deeper understanding of the topic and emerges as a key strategy for advancing programming learning. The automatic summarization was implemented by adapting a technique known as TextRank8 to the domain of the Mumuki forum in Spanish.

Recent advancements in multimodal foundation models have showcased impressive capabilities in understanding and reasoning with visual and textual information. Adapting these foundation models trained for general usage to specialized domains like biomedicine requires large-scale domain-specific instruction datasets. While existing works have explored curating such datasets automatically, the resultant datasets are not explicitly aligned with domain expertise. In this work, we propose a data-centric framework, Biomedical Visual Instruction Tuning with Clinician Preference Alignment (BioMed-VITAL), that incorporates clinician preferences into both stages of generating and selecting instruction data for tuning biomedical multimodal foundation models. First, during the generation stage, we prompt the GPT-4V generator with a diverse set of clinician-selected demonstrations for preference-aligned data candidate generation. Then, during the selection phase, we train a separate selection model, which explicitly distills clinician and policy-guided model preferences into a rating function to select high-quality data for medical instruction tuning. Results show that the model tuned with the instruction-following data from our method demonstrates a significant improvement in open visual chat (18.5% relatively) and medical VQA (win rate up to 81.73%). Our instruction-following data and models are available at BioMed-VITAL.github.io.

The definition of Data Science is a hotly debated topic. For many, the definition is a simple shortcut to Artificial Intelligence or Machine Learning. However, there is far more depth and nuance to the field of Data Science than a simple shortcut can provide. The School of Data Science at the University of Virginia has developed a novel model for the definition of Data Science. This model is based on identifying a unified understanding of the data work done across all areas of Data Science. It represents a generational leap forward in how we understand and teach Data Science. In this paper we will present the core features of the model and explain how it unifies various concepts going far beyond the analytics component of AI. From this foundation we will present our Undergraduate Major curriculum in Data Science and demonstrate how it prepares students to be well-rounded Data Science team members and leaders. The paper will conclude with an in-depth overview of the Foundations of Data Science course designed to introduce students to the field while also implementing proven STEM oriented pedagogical methods. These include, for example, specifications grading, active learning lectures, guest lectures from industry experts and weekly gamification labs.

A default assumption in the design of reinforcement-learning algorithms is that a decision-making agent always explores to learn optimal behavior. In sufficiently complex environments that approach the vastness and scale of the real world, however, attaining optimal performance may in fact be an entirely intractable endeavor and an agent may seldom find itself in a position to complete the requisite exploration for identifying an optimal policy. Recent work has leveraged tools from information theory to design agents that deliberately forgo optimal solutions in favor of sufficiently-satisfying or satisficing solutions, obtained through lossy compression. Notably, such agents may employ fundamentally different exploratory decisions to learn satisficing behaviors more efficiently than optimal ones that are more data intensive. While supported by a rigorous corroborating theory, the underlying algorithm relies on model-based planning, drastically limiting the compatibility of these ideas with function approximation and high-dimensional observations. In this work, we remedy this issue by extending an agent that directly represents uncertainty over the optimal value function allowing it to both bypass the need for model-based planning and to learn satisficing policies. We provide simple yet illustrative experiments that demonstrate how our algorithm enables deep reinforcement-learning agents to achieve satisficing behaviors. In keeping with previous work on this setting for multi-armed bandits, we additionally find that our algorithm is capable of synthesizing optimal behaviors, when feasible, more efficiently than its non-information-theoretic counterpart.

Online continual learning (OCL) involves deep neural networks retaining knowledge from old data while adapting to new data, which is accessible only once. A critical challenge in OCL is catastrophic forgetting, reflected in reduced model performance on old data. Existing replay-based methods mitigate forgetting by replaying buffered samples from old data and learning current samples of new data. In this work, we dissect existing methods and empirically discover that learning and replaying in the same feature space is not conducive to addressing the forgetting issue. Since the learned features associated with old data are readily changed by the features related to new data due to data imbalance, leading to the forgetting problem. Based on this observation, we intuitively explore learning and replaying in different feature spaces. Learning in a feature subspace is sufficient to capture novel knowledge from new data while replaying in a larger feature space provides more feature space to maintain historical knowledge from old data. To this end, we propose a novel OCL approach called experience replay with feature subspace learning (ER-FSL). Firstly, ER-FSL divides the entire feature space into multiple subspaces, with each subspace used to learn current samples. Moreover, it introduces a subspace reuse mechanism to address situations where no blank subspaces exist. Secondly, ER-FSL replays previous samples using an accumulated space comprising all learned subspaces. Extensive experiments on three datasets demonstrate the superiority of ER-FSL over various state-of-the-art methods.

With the advent of large pre-trained transformer models, fine-tuning these models for various downstream tasks is a critical problem. Paucity of training data, the existence of data silos, and stringent privacy constraints exacerbate this fine-tuning problem in the medical imaging domain, creating a strong need for algorithms that enable collaborative fine-tuning of pre-trained models. Moreover, the large size of these models necessitates the use of parameter-efficient fine-tuning (PEFT) to reduce the communication burden in federated learning. In this work, we systematically investigate various federated PEFT strategies for adapting a Vision Transformer (ViT) model (pre-trained on a large natural image dataset) for medical image classification. Apart from evaluating known PEFT techniques, we introduce new federated variants of PEFT algorithms such as visual prompt tuning (VPT), low-rank decomposition of visual prompts, stochastic block attention fine-tuning, and hybrid PEFT methods like low-rank adaptation (LoRA)+VPT. Moreover, we perform a thorough empirical analysis to identify the optimal PEFT method for the federated setting and understand the impact of data distribution on federated PEFT, especially for out-of-domain (OOD) and non-IID data. The key insight of this study is that while most federated PEFT methods work well for in-domain transfer, there is a substantial accuracy vs. efficiency trade-off when dealing with OOD and non-IID scenarios, which is commonly the case in medical imaging. Specifically, every order of magnitude reduction in fine-tuned/exchanged parameters can lead to a 4% drop in accuracy. Thus, the initial model choice is crucial for federated PEFT. It is preferable to use medical foundation models learned from in-domain medical image data (if available) rather than general vision models.