Predicting match outcomes typically depends on rich performance statistics. This work investigates whether it is possible to extract team synergy from only team composition and use it for sports outcome prediction. We model each player as a token and use a Siamese transformer to learn within‑team interactions and compare team representations to predict the win probability. Across multiple datasets, the model consistently outperforms classical baselines such as multilayer perceptrons and non‑Siamese transformers. Results suggest that synergy between teammates can be captured from team composition alone. When optional auxiliary information is available, a parallel Siamese branch yields additional performance gains. Overall, this study shows that team composition alone contains meaningful predictive structure that modern attention‑based models can effectively extract.

Recent advances in machine learning (ML) have enabled AI applications in mental disorder diagnosis, but many methods remain black-box or rely on post-hoc explanations which are not straightforward or actionable for mental health practitioners. Meanwhile, interpretable methods, such as k-nearest neighbors (k-NN) classification, struggle with complex or high-dimensional data. Moreover, there is a lack of study on users' real experience with interpretable AI. This study demonstrates a network-based k-NN model (NN-kNN) that combines the interpretability with the predictive power of neural networks. The model prediction can be fully explained in terms of activated features and neighboring cases. We experimented with the model to predict the risks of depression and interviewed practitioners in a qualitative study. The feedback of the practitioners emphasized the model's adaptability, integration of clinical expertise, and transparency in the diagnostic process, highlighting its potential to ethically improve the diagnostic precision and confidence of the practitioner.

This short paper presents preliminary research on the Case-Enhanced Vision Transformer (CEViT), a similarity measurement method aimed at improving the explainability of similarity assessments for image data. Initial experimental results suggest that integrating CEViT into k-Nearest Neighbor (k-NN) classification yields classification accuracy comparable to state-of-the-art computer vision models, while adding capabilities for illustrating differences between classes. CEViT explanations can be influenced by prior cases, to illustrate aspects of similarity relevant to those cases.

Recent research brings the strengths of neural networks to bear on CBR tasks such as similarity assessment and case adaptation. This paper further advances this direction by implementing both retrieval and adaptation as a single neural network. Such an approach has multiple goals: From the perspective of CBR, it enables harmonizing the interaction between feature extraction, retrieval/similarity assessment, and case adaptation through end-to-end training. From the perspective of neural networks, a neural network implementing CBR processes ceases to be a black box and provides the natural interpretability of CBR. As a first step towards this goal, this paper presents neural network based k-nearest neighbor (NN-kNN), a network architecture that can be interpreted as a k-NN method. Unlike other network architectures, NN-kNN’s decisions can be fully explained in terms of surface features, feature/case weights and nearest neighbors. It can be trained or fine-tuned using existing neural network methods. This study illustrates its feasibility and examines its strengths and limitations. The approach is evaluated for classification and regression tasks comparing NN-kNN, a standard neural network, and k-NN models using state-of-the-art distance metric learning algorithms. In these tests, NN-kNN achieves equal or less error when compared to the other models, while being fully interpretable as a k-NN method. The study also considered the limitations of NN-kNN and future directions to alleviate them.

Integrations of case-based reasoning (CBR) with neural approaches are appealing because of their complimentary characteristics. This chapter presents research on neuro-symbolic integrations to support CBR, to reduce knowledge engineering and improve performance for CBR systems. It summarizes three strands of research: First, on extracting features for case retrieval from deep neural networks to use in concert with expert-generated features, second, on applying neural networks to learn to adapt the solutions of retrieved cases to fit new situations, and third, on harmonizing similarity learning with case adaptation learning, in order to focus retrieval on adaptable cases. It summarizes strengths, weaknesses and tradeoffs of these approaches, and points to future challenges for neuro-CBR integrations.

Generating semi-factual and counterfactual explanations from images requires methods for extracting and adjusting appropriate image features. This short paper presents initial research on a counterfactual generation method for images based on class-to-class variational autoencoders (C2C-VAEs). Initial experiments illustrate substantial speed increase in counterfactual generation while suggesting that the method achieves reasonable counterfactual quality compared to the state of the art. The paper closes by discussing tradeoffs of the approach.

Neural network architectures for case adaptation in case-based reasoning (CBR) have received considerable attention. However, architectural gaps and general questions remain. First, existing architectures focus on adaptation of numeric attributes alone. Second, some proposed neural network adaptation architectures operate directly on pairs of cases, so could be performing direct prediction instead of adaptation. Third, it is unclear how the effectiveness of CBR systems with neural network components compares to that of networks alone. This paper addresses these questions. It extends a neural network-based case difference heuristic (NN-CDH) approach to handle both numeric and nominal attributes, in an architecture that applies to both regression and classification domains. The network predicts solution difference based on problem difference, ensuring that it learns adaptations. The paper presents experiments for both classification and regression tasks that compare performance of a neural network to a baseline CBR system and CBR variants with different retrieval schemes and adaptation schemes, on both real data and controlled artificial data sets. In these tests, CBR with the extended NN-CDH generally performs comparably to the baseline neural network, and NN-CDH consistently improves the results from naive retrieval but may worsen the results of network-based retrieval.

Case-based reasoning (CBR) research has developed numerous methods for learning to improve case retrieval and adaptation knowledge. Learning for each type of knowledge is usually pursued independently. However, it is well known that the knowledge containers of CBR are tightly coupled, in that changes in one can affect requirements for another, which suggests potential benefit for coupling learning across knowledge containers. This paper proposes applying alternative optimization to learn retrieval and adaptation knowledge together, in order to harmonize their behaviors. For a testbed system using neural network based similarity and adaptation, this study compares alternative optimization, independent learning, and learning by prioritizing adaptation for adaptation-guided retrieval. Results support that alternative optimization can help to balance both components and achieve good performance.

Case-based Reasoning (CBR) solves a new problem by retrieving a stored case for a similar problem and adapting its solution to fit. Acquiring the required case adaptation knowledge is a classic problem. A popular method for addressing it is the case difference heuristic (CDH) approach, which learns adaptations from pairs of cases based on their problem differences and solution differences. The CDH approach was originally used to generate adaptation rules, but recent CBR research on case-based regression has investigated replacing learning rules with learning CDH-based network models for adaptation. This paper presents and evaluates a neural network-based CDH approach for learning adaptation models for classification, C-NN-CDH. It examines three variants, (1) training a single neural network on problem-solution differences, (2) segmenting adaptation knowledge by the classes of source cases, with a separate neural network to generate adaptations for each class, and (3) adapting from an ensemble of source cases and taking the majority vote. Experimental results demonstrate improved performance compared to previous research on statistical methods for computing CDH differences for classification. Additional results support that C-NN-CDH achieves classification performance comparable to that of multiple classic classification approaches.

The case difference heuristic (CDH) approach is a knowledge-light method for learning case adaptation knowledge from the case base of a case-based reasoning system. Given a pair of cases, the CDH approach attributes the difference in their solutions to the difference in the problems they solve, and generates adaptation rules to adjust solutions accordingly when a retrieved case and new query have similar problem differences. As an alternative to learning adaptation rules, several researchers have applied neural networks to learn to predict solution differences from problem differences. Previous work on such approaches has assumed that the feature set describing problems is predefined. This paper investigates a two-phase process combining deep learning for feature extraction and neural network based adaptation learning from extracted features. Its performance is demonstrated in a regression task on an image data: predicting age given the image of a face. Results show that the combined process can successfully learn adaptation knowledge applicable to nonsymbolic differences in cases. The CBR system achieves slightly lower performance overall than a baseline deep network regressor, but better performance than the baseline on novel queries.

Case-based reasoning (CBR) is a knowledge-based reasoning and learning methodology that applies prior cases—records of prior instances or experiences—by adapting their lessons to solve new problems. The CBR process enables explainable reasoning from few examples, with minimal learning cost. However, the success of CBR depends on having appropriate similarity and adaptation knowledge, which may be hard to acquire. This paper illustrates the opportunity to leverage neural network methods to reduce the knowledge engineering burden for case-based reasoning. It presents an experimental example from ongoing work on refining the case difference heuristic approach to learning case adaptation knowledge by applying neural network learning.

Many degree-seeking college students struggle academically and ultimately never graduate. Academic challenges and persistence within the major are especially salient issues for students who major in science, technology, engineering, and mathematics. Academic probation serves as a means for informing students that they are at risk of dismissal, and many colleges and universities offer services to help students placed on probation to succeed academically. This paper presents two studies that examined the effectiveness of a goal-setting academic advising intervention for improving the grades of engineering students who were on academic probation; one study used a regression discontinuity design, and the other used an experimental design. The findings of both studies support the same overall conclusion: The intervention notably increased the grades of engineering students on probation who are beyond their first year of college, but it was not effective for students in their first year. This brief academic enhancement intervention appears to constitute a cost-effective strategy for bolstering the academic success of at-risk college students after their first year.

Case-based classification is normally based on similarity between a query and class members in the case base. This paper proposes a difference-based approach, class-to-class siamese network (C2C-SN) classification, in which classification is based on learning patterns of both similarity and difference between classes. A C2C-SN learns patterns from one class C_i to another class C_j . The network can then be used, given two cases, to determine whether their similarity and difference conform to the learned patterns. If they do, it provides evidence for their belonging to the corresponding classes. We demonstrate the use of C2C-SNs for classification, explanation, and prototypical case finding. We demonstrate that C2C-SN classification can achieve good accuracy for case pairs, with the benefit of one-shot learning inherited from siamese networks.

The ability of case-based reasoning systems to deal with new problems depends on the effectiveness of their case adaptation. One approach to increasing flexibility for novel problems is to perform adaptations by using adaptation paths—chains of adaptations—to address differences beyond those addressable by applying single adaptation rules. A recent approach to adaptation path generation, ROAD, proposes building adaptation paths using heuristic search guided by similarity, with a “reset” mechanism for recovering when similarity fails to predict adaptability. The ROAD approach is beneficial when similarity and adaptability are well aligned, but can make poor choices when similarity and adaptability diverge, increasing adaptation cost. This paper presents methods for increasing adaptation efficiency by maintenance exploiting information from adaptation path generation. The methods improve the similarity measure to better reflect adaptability and condense the adaptation rule set. Experimental evaluation supports the benefits for improving adaptation efficiency while preserving accuracy.

The case adaptation process in case-based reasoning is often modeled as having two steps: enumerating differences between a new problem and the problem part of a retrieved case and then applying an adaptation rule for each difference. This model is sufficient when (1) predefined adaptation rules exist for all differences the system encounters, and (2) adaptation rules are sufficiently independent that interactions are not a major issue. This paper presents an approach to handling case adaptation when these assumptions fail. It proposes an approach, RObust ADaptation (ROAD), that uses heuristics to guide multi-step adaptations, with each adaptation chosen in the context of adaptations applied previously. To reduce the potential for accumulated degradation of solution quality from long adaptation chains, it performs incremental retrieval of new source cases along the adaptation path, resetting the partially modified case to the "ground truth" of existing cases when an existing case is nearby. An evaluation supports the benefits of the model and illuminates some tradeoffs.

Traditional case-based classification methods are based on feature similarity. In contrast, class-to-class (C2C) weighting also considers whether the difference between two cases has been seen before. Combined with instance-specific weighting, C2C weighting learns the local patterns of both simi- larities and differences (shortened as patterns ). Once C2C weightings has learned the pattern between case A of class C 1 and some set of cases R of class C 2 , given a query Q whose difference from A matches the pattern between A and R , then we can skip cases around A and continue the search for near neighbors around R . Based on this, we developed an algorithm, C2C trace re- trieval, which quickly traverses promising cases, retrieves relevant cases from different classes, and provides an informed hypothesis of the query’s class. C2C trace retrieval achieves great efficiency at a reasonable cost of accuracy. Therefore, C2C trace retrieval can be used as a fast classification method or as the first pass for a more sophisticated method.

Over the past two decades, counseling psychology has emerged as an international discipline. Despite efforts to internationalize counseling psychology and increase cross-cultural relationships, few studies have considered international collaboration on scientific scholarship as a metric or factor of internationalization. Therefore, using social network analyses, we tested the occurrence and patterns of collaborations between authors from different countries on published studies in the Journal of Counseling Psychology and The Counseling Psychologist from 2005 to 2015. Results indicated that a small proportion of possible international collaborations existed, as measured by network density, and the majority of collaborations involved authors affiliated with institutions in the United States. However, international collaboration significantly increased over time, as measured by the density of the network. Our findings are consistent with previous research suggesting a lack of cross-national scholarship in counseling psychology and identify global areas for further cross-cultural development.