Set Of Experiments Research Articles

Revisiting Realistic Test-Time Training: Sequential Inference and Adaptation by Anchored Clustering Regularized Self-Training.

Deploying models on target domain data subject to distribution shift requires adaptation. Test-time training (TTT) emerges as a solution to this adaptation under a realistic scenario where access to full source domain data is not available, and instant inference on the target domain is required. Despite many efforts into TTT, there is a confusion over the experimental settings, thus leading to unfair comparisons. In this work, we first revisit TTT assumptions and categorize TTT protocols by two key factors, i.e., whether testing data is sequentially streamed and whether source model is allowed to be trained with modified loss function. Among the multiple protocols, we adopt a realistic sequential test-time training (sTTT) protocol, under which we develop a test-time anchored clustering (TTAC) approach to enable stronger test-time feature learning. TTAC discovers clusters in both source and target domains and matches the target clusters to the source ones to improve adaptation. When source domain information is strictly absent (i.e., source-free) we further develop an efficient method to infer source domain distributions for anchored clustering. Finally, self-training (ST) has demonstrated great success in learning from unlabeled data and we empirically figure out that applying ST alone to TTT is prone to confirmation bias. Therefore, a more effective TTT approach is introduced by regularizing self-training with anchored clustering, and the improved model is referred to as TTAC++. We demonstrate that, under all TTT protocols, TTAC++ consistently outperforms the state-of-the-art methods on five TTT datasets, including corrupted target domain, selected hard samples, synthetic-to-real adaptation and adversarially attacked target domain. We hope this work will provide a fair benchmarking of TTT methods, and future research should be compared within respective protocols.

On Network Structural and Temporal Encodings: A Space and Time Odyssey.

The dynamic network visualization design space consists of two major dimensions: network structural and temporal representation. As more techniques are developed and published, a clear need for evaluation and experimental comparisons between them emerges. Most studies explore the temporal dimension and diverse interaction techniques supporting the participants, focusing on a single structural representation. Empirical evidence about performance and preference for different visualization approaches is scattered over different studies, experimental settings, and tasks. This paper aims to comprehensively investigate the dynamic network visualization design space in two evaluations. First, a controlled study assessing participants' response times, accuracy, and preferences for different combinations of network structural and temporal representations on typical dynamic network exploration tasks, with and without the support of standard interaction methods. Second, the best-performing combinations from the first study are enhanced based on participants' feedback and evaluated in a heuristic-based qualitative study with visualization experts on a real-world network. Our results highlight node-link with animation and playback controls as the best-performing combination and the most preferred based on ratings. Matrices achieve similar performance to node-link in the first study but have considerably lower scores in our second evaluation. Similarly, juxtaposition exhibits evident scalability issues in more realistic analysis contexts.

Perceptron Theory Can Predict the Accuracy of Neural Networks.

Multilayer neural networks set the current state of the art for many technical classification problems. But, these networks are still, essentially, black boxes in terms of analyzing them and predicting their performance. Here, we develop a statistical theory for the one-layer perceptron and show that it can predict performances of a surprisingly large variety of neural networks with different architectures. A general theory of classification with perceptrons is developed by generalizing an existing theory for analyzing reservoir computing models and connectionist models for symbolic reasoning known as vector symbolic architectures. Our statistical theory offers three formulas leveraging the signal statistics with increasing detail. The formulas are analytically intractable, but can be evaluated numerically. The description level that captures maximum details requires stochastic sampling methods. Depending on the network model, the simpler formulas already yield high prediction accuracy. The quality of the theory predictions is assessed in three experimental settings, a memorization task for echo state networks (ESNs) from reservoir computing literature, a collection of classification datasets for shallow randomly connected networks, and the ImageNet dataset for deep convolutional neural networks. We find that the second description level of the perceptron theory can predict the performance of types of ESNs, which could not be described previously. Furthermore, the theory can predict deep multilayer neural networks by being applied to their output layer. While other methods for prediction of neural networks performance commonly require to train an estimator model, the proposed theory requires only the first two moments of the distribution of the postsynaptic sums in the output neurons. Moreover, the perceptron theory compares favorably to other methods that do not rely on training an estimator model.

Creditors’ Role in Shaping Asymmetric Cost Behavior: Evidence from Debt Covenant Violation

ABSTRACT This study uses covenant violations as a quasi-natural experimental setting to examine creditors’ roles in shaping corporate cost behavior. Using a regression discontinuity design, I find that cost stickiness experiences a sharp decline following debt covenant violations when control rights are transferred to creditors. The cost stickiness effect is more substantial for borrowers with lower credit ratings and when creditors possess greater bargaining power. The effect is also more pronounced during industry downturns when borrowers have fewer alternative sources of finance. Results are consistent when I use alternative measures of cost stickiness and alternative research designs. Overall, my evidence indicates that creditors play a monitoring role in firms’ cost behaviors and identifies a specific channel (loan covenants) through which the misalignment of incentives can impact cost asymmetry. JEL Classifications: D24; G32; M41.

TMLNet: Triad Multitask Learning Network for multiobjective based change detection

Change detection is an essential computer vision task in remote sensing applications. It faces challenges of image registration errors, variation in image capturing conditions, clouds, etc. It is observed that an accurate change detection task requires effective feature learning, which needs a noiseless representation. Moreover, change detection improves with error-free image reconstruction as an auxiliary task, which needs joint feature representation by a feature extractor. In this work, we proposed a novel triad (which is a combination of input images and its difference) learning based multiresolution architecture TMLNet for effective change detection. We also proposed a multi-context local self-attention module to efficiently calculate long-range pixel relations with multiple contexts. An enhanced backbone module with top-down connections and multi-scale channel and spatial attention is utilized for change map generation. It provides less noisy features extracted through the backbone. Laplacian pyramid loss is used to preserve small details in feature reconstruction. A set of comprehensive experimentations reveals that the proposed scheme achieved the state-of-the-art result for the F1 Score, intersection over union, and overall accuracy values in seven benchmark datasets. Our model code is available at https://github.com/chouhan-avinash/TMLNet.

Domain-Independent Lifelong Problem Solving Through Distributed ALife Actors.

A domain-independent problem-solving system based on principles of Artificial Life is introduced. In this system, DIAS, the input and output dimensions of the domain are laid out in a spatial medium. A population of actors, each seeing only part of this medium, solves problems collectively in it. The process is independent of the domain and can be implemented through different kinds of actors. Through a set of experiments on various problem domains, DIAS is shown able to solve problems with different dimensionality and complexity, to require no hyperparameter tuning for new problems, and to exhibit lifelong learning, that is, to adapt rapidly to run-time changes in the problem domain, and to do it better than a standard, noncollective approach. DIAS therefore demonstrates a role for ALife in building scalable, general, and adaptive problem-solving systems.

Experimental Validation of a Command and Control Traffic Detection Model

Network intrusion detection systems (NIDS) are commonly used to detect malware communications, including command-and-control (C2) traffic from botnets. NIDS performance assessments have been studied for decades, but mathematical modeling has rarely been used to explore NIDS performance. This paper details a mathematical model that describes a NIDS performing packet inspection and its detection of malware's C2 traffic. The paper further describes an emulation testbed and a set of cyber experiments that used the testbed to validate the model. These experiments included a commonly used NIDS (Snort) and traffic with contents from a pervasive malware (Emotet). Results are presented for two scenarios: a nominal scenario and a “stressed” scenario in which the NIDS cannot process all incoming packets. Model and experiment results match well, with model estimates mostly falling within 95 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\%$</tex-math></inline-formula> confidence intervals on the experiment means. Model results were produced 70-3000 times faster than the experimental results. Consequently, the model's predictive capability could potentially be used to support decisions about NIDS configuration and effectiveness that require high confidence results, quantification of uncertainty, and exploration of large parameter spaces. Furthermore, the experiments provide an example for how emulation testbeds can be used to validate cyber models that include stochastic variability.

Hypergraph Isomorphism Computation.

The isomorphism problem, crucial in network analysis, involves analyzing both low-order and high-order structural information. Graph isomorphism algorithms focus on structural equivalence to simplify solver space, aiding applications like protein design, chemical pathways, and community detection. However, they fall short in capturing complex high-order relationships, unlike hypergraph isomorphism methods. Traditional hypergraph methods face challenges like high memory use and inaccurate identification, leading to poor performance. To overcome these, we introduce a hypergraph Weisfeiler-Lehman (WL) test algorithm, extending the WL test from graphs to hypergraphs, and develop a hypergraph WL kernel framework with two variants: the Hypergraph WL Subtree Kernel and Hypergraph WL Hyperedge Kernel. The Hypergraph WL Subtree Kernel counts different types of rooted subtrees and generates the final feature vector for a given hypergraph by comparing the number of different types of rooted subtrees. The Subtree Kernel identifies different rooted subtrees, while the Hyperedge Kernel focuses on hyperedges' vertex labels, enhancing feature vector generation. In order to fulfill our research objectives, a comprehensive set of experiments was meticulously designed, including seven graph classification datasets and 12 hypergraph classification datasets. Results on graph classification datasets indicate that the Hypergraph WL Subtree Kernel can achieve the same performance compared with the classical Graph Weisfeiler-Lehman Subtree Kernel. Results on hypergraph classification datasets show significant improvements compared to other typical kernel-based methods, which demonstrates the effectiveness of the proposed methods. In our evaluation, our proposed methods outperform the second-best method in terms of runtime, running over 80 times faster when handling complex hypergraph structures. This significant speed advantage highlights the great potential of our methods in real-world applications.

Formation Control of Multiple Autonomous Mobile Robots Using Turkish Natural Language Processing

People use natural language to express their thoughts and wishes. As robots reside in various human environments, such as homes, offices, and hospitals, the need for human–robot communication is increasing. One of the best ways to achieve this communication is the use of natural languages. Natural language processing (NLP) is the most important approach enabling robots to understand natural languages and improve human–robot interaction. Also, due to this need, the amount of research on NLP has increased considerably in recent years. In this study, commands were given to a multiple-mobile-robot system using the Turkish natural language, and the robots were required to fulfill these orders. Turkish is classified as an agglutinative language. In agglutinative languages, words combine different morphemes, each carrying a specific meaning, to create complex words. Turkish exhibits this characteristic by adding various suffixes to a root or base form to convey grammatical relationships, tense, aspect, mood, and other semantic nuances. Since the Turkish language has an agglutinative structure, it is very difficult to decode its sentence structure in a way that robots can understand. Parsing of a given command, path planning, path tracking, and formation control were carried out. In the path-planning phase, the A* algorithm was used to find the optimal path, and a PID controller was used to follow the generated path with minimum error. A leader–follower approach was used to control multiple robots. A platoon formation was chosen as the multi-robot formation. The proposed method was validated on a known map containing obstacles, demonstrating the system’s ability to navigate the robots to the desired locations while maintaining the specified formation. This study used Turtlebot3 robots within the Gazebo simulation environment, providing a controlled and replicable setting for comprehensive experimentation. The results affirm the feasibility and effectiveness of employing NLP techniques for the formation control of multiple mobile robots, offering a robust and effective method for further research and development on human–robot interaction.

A novel bilateral stream neural network for melt pool monitoring during laser direct energy deposition

Detection and classification methods for the melt pool state in laser direct energy deposition (L-DED) can significantly help predict defects and mechanical properties of L-DED metal parts. Although traditional machine learning algorithms based on physical modeling methods and convolutional neural networks have recently been introduced into melt pool state identification, these methods rely on complex artificially designed features or cannot simultaneously detect defects in multiple dimensions. In this paper, a novel bilateral stream neural network was designed for melt pool identification, which performs defect identification in two label dimensions simultaneously. Two sets of single-channel experiments were designed to collect the dataset captured by a high-speed camera. By cutting the metal parts and marking them with professional equipment operated by professionals, the dataset was labeled according to the bonding condition and dilution rate criteria. Without an additive model structure, the model achieved 95.2% accuracy in identifying defects in the bonding condition and 92.8% in determining deficiencies in the dilution rate. In order to explain the identification mechanism of the model, the CAM method was utilized for the visual display of the model recognition process, which provides a potential application solution for the online monitoring method of the L-DED.

Deformation characteristics and creep behaviour of rigid particulates-EPS beads composites

The compression behaviour of the mixture of glass beads (representing rigid particles) and EPS beads (representing deformable particles) during the loading-unloading process is systematically examined through performing two sets of large-size oedometer experiments, including incremental step-by-step and one-step loading scenarios. At each step during the loading-unloading cycle, the void ratio (e) and the at-rest coefficient of lateral earth pressure (K0) are measured for pure rigid samples and rigid-soft particle mixtures. To consider the creep effect, the overburden pressure at the final loading step is maintained on the sample for 24 h prior to unloading. The results show that at a given overburden pressure, with the addition of soft particles to the pure rigid aggregates, the values of e and K0 decrease. Additionally, for both pure rigid samples and rigid-soft particle mixtures, with increasing the overburden pressure, e decreases whereas K0 augments. Moreover, due to the creep behaviour during the constant loading step, K0 decreases over time for both samples; a phenomenon which is observed to be more pronounced for pure rigid aggregates compared to rigid-soft particle mixtures. Finally, a well-established creep model is used to simulate the creep behaviour of pure rigid samples and rigid-soft particle composites.

Deep Attentive Interest Collaborative Filtering for Recommender Systems

Collaborative filtering (CF) is a pivotal building block in commercial recommender systems due to its strength and utility in user interest modeling. Recently, many researchers have turned to deep learning as a way to capture richer collaborative signals from user-item feature interactions. However, most deep-based methods only consider nonlinear, high-order interactions while ignoring the explicit collaborative signals in low-order interactions. They also typically ignore the quality of the user and item profiles. These are cornerstones in item recommendation that, we argue, must be considered for high-quality recommendations. Hence, we propose Deep Attentive Interest Collaborative Filtering (DAICF) to overcome these limitations. DAICF profiles users based on their interactive items, i.e., user neighborhood information. Similarly, item profiles are based on users who had interacted with it, i.e., item neighborhood information. Given a user's profile varies over different items, DAICF accurately models his attentive interests based on the specific target item. Low-order collaborative signals are captured by a shallow component, and high-order collaborative signals are captured by a deep component. These two complementary collaborative signals are then fused to provide rich recommendations that cut through today's information overload. By designing a personalized feature extraction method based on bilateral neighborhood information to solve the data sparsity problem in recommender systems, DAICF can dynamically distinguish the importance of a user's historical interaction items for predicting user preferences for a specific target item. A set of experiments against four real-world datasets validate that DAICF outperforms the most recent state-of-the-art recommendation algorithms and justifies the effectiveness and interpretability of our method.

Multi-Behavior Graph Neural Networks for Recommender System.

Recommender systems have been demonstrated to be effective to meet user's personalized interests for many online services (e.g., E-commerce and online advertising platforms). Recent years have witnessed the emerging success of many deep-learning-based recommendation models for augmenting collaborative filtering (CF) architectures with various neural network architectures, such as multilayer perceptron and autoencoder. However, the majority of them model the user-item relationship with single type of interaction, while overlooking the diversity of user behaviors on interacting with items, which can be click, add-to-cart, tag-as-favorite, and purchase. Such various types of interaction behaviors have great potential in providing rich information for understanding the user preferences. In this article, we pay special attention on user-item relationships with the exploration of multityped user behaviors. Technically, we contribute a new multi-behavior graph neural network (MBRec), which specially accounts for diverse interaction patterns and the underlying cross-type behavior interdependencies. In the MBRec framework, we develop a graph-structured learning framework to perform expressive modeling of high-order connectivity in behavior-aware user-item interaction graph. After that, a mutual relationship encoder is proposed to adaptively uncover complex relational structures and make aggregations across layer-specific behavior representations. Through comprehensive evaluation on real-world datasets, the advantages of our MBRec method have been validated under different experimental settings. Further analysis verifies the positive effects of incorporating the multi-behavioral context into the recommendation paradigm. In addition, the conducted case studies offer insights into the interpretability of user multi-behavior representations. We release our model implementation at https://github.com/akaxlh/MBRec.

Out-of-Sample Tuning for Causal Discovery.

Causal discovery is continually being enriched with new algorithms for learning causal graphical probabilistic models. Each one of them requires a set of hyperparameters, creating a great number of combinations. Given that the true graph is unknown and the learning task is unsupervised, the challenge to a practitioner is how to tune these choices. We propose out-of-sample causal tuning (OCT) that aims to select an optimal combination. The method treats a causal model as a set of predictive models and uses out-of-sample protocols for supervised methods. This approach can handle general settings like latent confounders and nonlinear relationships. The method uses an information-theoretic approach to be able to generalize to mixed data types and a penalty for dense graphs to penalize for complexity. To evaluate OCT, we introduce a causal-based simulation method to create datasets that mimic the properties of real-world problems. We evaluate OCT against two other tuning approaches, based on stability and in-sample fitting. We show that OCT performs well in many experimental settings and it is an effective tuning method for causal discovery.

Pa-Count: Passenger Counting in Vehicles Using Wi-Fi Signals

Passenger counting is crucial for many applications such as vehicle scheduling and traffic capacity assessment. However, most of the existing solutions are either high-cost, privacy invasive or not suitable for passengers the vehicle scenarios. In this work, we propose the <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Pa</u> - <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Count</u> , an effective real-time <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Pa</u> ssenger <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Count</u> ing system deployed inside the vehicle via using Wi-Fi <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CSI</i> (Channel State Information). Specifically, in Pa-Count, we design a set of combined filters to eliminate environmental interference and enhance CSI quality. In so doing, we can identify the fluctuation of weak CSI caused by passengers' subtle movement, i.e., the fidgeting, and then obtain the distribution of fidgeting period and silent period. Following that, we describe the subtle movements of passengers via power law with exponential cutoff distribution and establish a counting model based on the queuing theory. A mathematical inference method with a priori probability is devised to calculate the number of real-time passengers through CSI. We evaluate the performance of the Pa-Count by conducting a set of experiments in real-world vehicle scenarios (including private car and subway). Experimental results show that Pa-Count can achieve robust performance with an average accuracy of over 92 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\%$</tex-math></inline-formula> .

Current Edge-Control Strategy for Multiple Energy Routers Based on Cyber-Energy Dual Modulations

Recently, energy routers have been widely applied in the dc microgrid to achieve the power balance among buses, and the relative control strategies with communication network assumption have been proposed. However, the communication network assumption cannot be satisfied in remote mountainous area and island area. To this end, this paper proposes one communication-free edge-control strategy for energy routers based on cyber-energy dual modulation to achieve power balance without communication network assumption. First, the network structure and system model are introduced, which is an fundamental pretreatment for the subsequent controller design. Second, the cyber-energy dual modulation approach without additional communication line is designed to exchange the information among distributed generations (DGs), which is an indispensable communication environment design for the subsequent controller design. Considering the system model feature and communication environment feature, the periodic dynamic event-triggered-based edge-control strategy is proposed in this paper. Finally, the proposed strategy effectiveness of the energy router is well verified through DC-DC converter experiment set and OPAL-RT test system.

Research on the Remaining Life Prediction Method of Rolling Bearings Based on Optimized TPA-LSTM

The whole life cycle degradation data set of rolling bearings has the characteristics of large capacity, diversity, and non-stationarity. As a powerful tool for processing such time series data in deep learning algorithms, LSTM is prone to the loss of important time series information in the process of the life prediction of rolling bearings, which leads to a decline in prediction accuracy. Therefore, a method for predicting the remaining useful life (RUL) of rolling bearings based on the combination of temporal pattern attention mechanism (TPA) and LSTM is proposed. The method firstly combines hierarchical clustering and principal component analysis (PCA) to construct a multi-faceted and multi-scale preferred feature set reflecting the degradation information of rolling bearings, then strengthens the information correlation between hidden layers of the LSTM model through TPA and optimates the parameters of the fusion model of TPA and LSTM by using the gazelle optimization algorithm (GOA). Finally, the model is applied to the experimental data set of rolling bearing degradation. The results show that, compared with the traditional model, this method is more suitable for the remaining life prediction of rolling bearings.

A Game-Theoretical Self-Adaptation Framework for Securing Software-Intensive Systems

Security attacks present unique challenges to the design of self-adaptation mechanism for software-intensive systems due to the adversarial nature of the environment. Game-theoretical approaches have been explored in security to model malicious behaviors and design reliable defense for the system in a mathematically grounded manner. However, modeling the system as a single player, as done in prior works, is insufficient for the system under partial compromise and for the design of fine-grained defensive policies where the rest of the system with autonomy can cooperate to mitigate the impact of attacks. To address such issues, we propose a new self-adaptation framework incorporating Bayesian game theory and model the defender (i.e., the system) at the granularity of components. Under security attacks, the architecture model of the system is automatically translated, by the proposed translation process with designed algorithms, into a multi-player Bayesian game. This representation allows each component to be modelled as an independent player, while security attacks are encoded as variant types for the components. By solving for pure equilibrium (i.e., adaptation response), the system’s optimal defensive strategy is dynamically computed, enhancing system resilience against security attacks by maximizing system utility. We validate the effectiveness of our framework through two sets of experiments using generic benchmark tasks tailored for the security domain. Additionally, we exemplify the practical application of our approach through a real-world implementation in the Secure Water Treatment System to demonstrates the applicability and potency in mitigating security risks.

Dynamic Path Planning Based on 3D Cloud Recognition for an Assistive Bathing Robot

Assistive bathing robots have become a popular point due to their metrics, such as a humanoid working approach in the solution of elder care. However, the abilities of dynamic recognition and path planning are the key to obtain the advantages. This paper proposes a novel approach to recognize and track the dynamical human back, and path planning on it via a 3D point cloud. Firstly, the human back geometric features are recognized through coarse-to-fine alignment. The Intrinsic Shape Signature (ISS) algorithm combined with the Fast Point Feature Histogram (FPFH) and the Sample Consensus Initial Alignment (SAC-IA) algorithm are adopted to complete the coarse alignment, and the Iterative Closest Point (ICP) algorithm is applied to the fine alignment to improve the accuracy of recognition. Then, the dynamic transformation matrix between the contiguous recognized back is deduced based on spatial motion between two adjacent recognized back point clouds. The path can be planned on the tracked human back. Finally, a set of testing experiments are conducted to verify the proposed algorithm. The results show that the running time is reduced by 66.18% and 96.29% compared with the other two common algorithms, respectively.

Abstract 4091: A novel cancer vaccine based on hyaluronic acid nanogel combined with adoptive T cell therapy induces complete regression of established tumors and long-lasting memory CD8+ T cells

Abstract Background: We have developed nanosized hydrogel particles (nanogels) to create new nanomaterials for biomedical applications. In particular, Hyaluronic Acid, partially hydrophobized by chemical modification with cholesteryl groups, has a distinguished characteristic to form physically cross-linked NanoGel particles. HANG efficiently forms a stable complex with an antigenic polypeptide (HANG-Vaccine) through hydrophobic interactions. In this study, we investigated the efficacy of HANG-V against tumors resistant to immune checkpoint inhibitors (ICIs) and adoptive TCR-T therapy. Methods: C57BL/6-derived B16F10 melanoma and BALB/c-derived CMS5a sarcoma are both known to be resistant to ICIs and adoptive TCR-T therapy. The efficacy of HANG-V was tested by using each tumor-bearing mice treatment model. HANG-V was prepared by mixing gp100 or CMS5a neoantigen (Ag) with HANG, respectively. Results: HANG-V was subcutaneously injected into B16F10-bearing C57BL/6 on days 7, 11, and 15, and CD8+ T cells from Pmel-1 mice were intravenously administered on days 8 and 12. Surprisingly, complete regression of established B16F10 tumors with more than 100 mm2 were observed. Vaccinated mice survived with changing hair color of mice into white, while non-vaccinated mice were all dead by day 25. Moreover, gp100-specific CTLs was observed abundantly in systemic including tumor site. Even two months after the treatment, transferred T cells with memory phenotype were persistently detected in peripheral bloods. Furthermore, booster vaccination elicited robust expansion of these memory T cells persisting for more than one year. In similar experimental settings, we treated CMS5a tumor-bearing mice by HANG-V combined with adoptive transfer of CD8+ T cells from DUC18 mice, whose T cells are specific for mutated ERK2 neoAg. All mice survived as well with complete reduction of established CMS5a tumors, rejected a rechallenge with CMS5a and induced an Ag-specific CTL responses on more than 2 years. Notably, by using neoantigen-knockout CMS5a, we obtained results suggesting that HANG-V with TCR-T therapy potently induced the so-called “antigen spreading”. Furthermore, induction of Ag-specific CTLs was associated with CD44 hyaluronic acid receptor which expressed in APCs crucial for cross-presentation to CD8+ T cells. Conclusions: HANG-V strongly enhanced the efficacy of adoptive TCR-T cell therapy against ICI refractory tumors leading to total tumor suppression. Furthermore, HANG-V induced potent and persistent Ag-specific CTLs systemically through interaction with CD44 hyaluronic receptor, allowing long-term protection of tumor recurrence with memory CD8+ T cells. Our studies may propose crucial insights for clinical application of HANG vaccine with adoptive T cell therapy in patient with ICI-resistant tumors with poor prognosis. Citation Format: Fumiyasu Momose, Takashi Nakai, Kohei Yabuuchi, Makiko Yamane, Tae Hayashi, Linan Wang, Yoshiyuki Nakagawa, Shogo Aso, Toru Katsumata, Tsuyoshi Shimoboji, Yoshihiro Miyahara. A novel cancer vaccine based on hyaluronic acid nanogel combined with adoptive T cell therapy induces complete regression of established tumors and long-lasting memory CD8+ T cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4091.