Interactive Learning Approach Research Articles

Use of interactive teaching –Learning approaches in anatomy

Introduction: Understanding of anatomy is fundamental for clinical practice; Anatomy must be taught and learnt within a context that is clinically meaningful and related to the competencies required by new medical graduates so that students understand its relevance to their future practice. Hence in this study we decided to implement interactive clinical anatomy lecture. Aim: To assess the impact of the interactive teaching learning methods on student cognitive outcomes and to evaluate the perceptions of students about interactive teaching learning approach. Materials and Methods: Faculty members were sensitized. The study was approved by institutional Ethics Committee. Written informed consent was taken from students. The anatomy knowledge was evaluated by a pre-test and post-test (10 MCQ), which were designed based on the learning objectives of the sessions. Student’s perception was evaluated through prevalidated questionnaires. The topic was chosen for interactive sessions namely “Venous drainage of lower limb” and we were including video clips and demonstration of Trendelenburg test in this session. Students had learned this topic already during didactic lecture as a part of routine timetable. The pre-test and post-test results of the study group were compared using ‘paired t’ test. Students feedback was determined by analysis of questionnaires based on 5 point Likert scale. Results: Total participants were 150. The mean score of pretest and posttest were 3.7 and 4.2 respectively. Standard deviation for pretest and posttest were 1.52 and 1.51 respectively. t value of this study was 7.99 and P value was 0.0000. Here P value is less than 0.001, it shows the highly significant difference between the mean of pretest and posttest score. It indicates definite improvement in the knowledge of students. In the prevalidated feedback questionnaires, 64.7% students had agreed that session was better than didactic lecture and 82.7% students had agreed that it was m

A glass-box interactive machine learning approach for solving NP-hard problems with the human-in-the-loop

The ultimate goal of the Machine Learning (ML) community is to develop algorithms that can automatically learn from data, to extract knowledge and to make decisions without any human intervention. Specifically, automatic Machine Learning (aML) approaches show impressive success, e.g. in speech/image recognition or autonomous drive and smart car industry. Recent results even demonstrate intriguingly that deep learning applied for automatic classification of skin lesions is on par with the performance of dermatologists, yet outperforms the average human efficiency. As human perception is inherently limited to 3D environments, such approaches can discover patterns, e.g. that two objects are similar, in arbitrarily high-dimensional spaces what no human is able to do. Humans can deal simultaneously only with limited amounts of data, whilst “big data” is not only beneficial but necessary for aML. However, in health informatics, there are few data sets; aML approaches often suffer from insufficient training samples. Many problems are computationally hard, e.g. subspace clustering, k-anonymization, or protein folding. Here, interactive machine learning (iML) could be successfully used, as a human-in-the-loop contributes to reduce a huge search space through heuristic selection of suitable samples. This can reduce the complexity of NP-hard problems through the knowledge brought in by a human agent involved into the learning algorithm. A huge motivation for iML is that standard black-box approaches lack transparency, hence do not foster trust and acceptance of ML among end-users. Most of all, rising legal and privacy aspects, e.g. the European General Data Protection Regulations (GDPR) make black-box approaches difficult to use, because they often are not able to explain why a decision has been made, e.g. why two objects are similar. All these reasons motivate the idea to open the black-box to a glass-box. In this paper, we present some experiments to demonstrate the effectiveness of the iML human-in-the-loop model, in particular when using a glass-box instead of a black-box model and thus enabling a human directly to interact with a learning algorithm. We selected the Ant Colony System (ACS) algorithm, and applied it on the Traveling Salesman Problem (TSP). The TSP-problem is a good example, because it is of high relevance for health informatics as for example on protein folding problem, thus of enormous importance for fostering cancer research. Finally, from studies of learning from observation, i.e. of how humans extract so much from so little data, fundamental ML-research also may benefit.

Extracting safety information from multi-lingual accident reports using an ontology-based approach

This paper describes an approach to extract meaning from multi-lingual free-text safety incident reports. A sample of 5065 safety incident reports from the Swiss Federal Office of Transport were used in the study. Each report was written in either German, French or Italian natural language. An interactive learning approach between a human and computer software was undertaken to identify key terms in the text that are relevant to discovering meaning. A multi-lingual ontology was created to join meaningful semantic patterns and identify specific classes of safety incident on the railway, including injuries occurring whilst passengers were boarding or alighting from vehicles, falling down stairs, struck by closing doors, or struck by objects such as suitcases. A graph database was used to query the text records via the ontology and identify reports of incidents in each class, regardless of the language used in the report. Fluent speakers of each language – German, French and Italian – reviewed the results to confirm true positive results and detect false positives. The performance of the process varied across languages and incident types, however the overall true positive rate was determined by the fluent speakers to be 98.5%.

Developing faculty leadership from ‘within’: a 12-year reflection from an internal faculty leadership development program of an academic health sciences center

ABSTRACTMost academic health sciences centers offer faculty leadership development programs (LDPs); however, the outcomes of LDPs are largely unknown. This article describes perspectives from our 12-year experience cultivating a formal faculty LDP within an academic health center and longitudinal outcomes of our LDP. Responding to faculty concerns from University of California San Francisco’s (UCSF) 2001 Faculty Climate Survey, UCSF established the UCSF-Coro Faculty Leadership Collaborative (FLC) in 2005. The FLC focused on building leadership skills using a cohort-based, experiential, interactive and collaborative learning approach. From 2005 to 2012, FLC has conducted training for 136 graduates over 7 cohorts with 97.6% completion rate. FLC faculty participants included 64% women and 13% underrepresented minority (URM). The proportions of graduates attaining leadership positions within UCSF such as deans or department chairs among all, URM, and women URM graduates were 9.6%, 33.3% and 45.5%, respectively. A 2013 online survey assessed 2005–2012 graduates’ perceived impacts from 8 months to 8 years after program completion and showed 91.7% of survey respondents felt the program both increased their understanding of UCSF as an organization and demonstrated the University’s commitment to foster faculty development. Qualitative results indicated that graduates perceived benefits at individual, interpersonal, and organizational levels. Though we did not directly assess impact on faculty recruitment and retention, the findings to date support cohort-based experiential learning in faculty leadership training development.

Interactive machine learning: experimental evidence for the human in the algorithmic loop

Recent advances in automatic machine learning (aML) allow solving problems without any human intervention. However, sometimes a human-in-the-loop can be beneficial in solving computationally hard problems. In this paper we provide new experimental insights on how we can improve computational intelligence by complementing it with human intelligence in an interactive machine learning approach (iML). For this purpose, we used the Ant Colony Optimization (ACO) framework, because this fosters multi-agent approaches with human agents in the loop. We propose unification between the human intelligence and interaction skills and the computational power of an artificial system. The ACO framework is used on a case study solving the Traveling Salesman Problem, because of its many practical implications, e.g. in the medical domain. We used ACO due to the fact that it is one of the best algorithms used in many applied intelligence problems. For the evaluation we used gamification, i.e. we implemented a snake-like game called Traveling Snakesman with the MAX–MIN Ant System (MMAS) in the background. We extended the MMAS–Algorithm in a way, that the human can directly interact and influence the ants. This is done by “traveling” with the snake across the graph. Each time the human travels over an ant, the current pheromone value of the edge is multiplied by 5. This manipulation has an impact on the ant’s behavior (the probability that this edge is taken by the ant increases). The results show that the humans performing one tour through the graphs have a significant impact on the shortest path found by the MMAS. Consequently, our experiment demonstrates that in our case human intelligence can positively influence machine intelligence. To the best of our knowledge this is the first study of this kind.

1323. Implementation and Effectiveness of the Asia Pacific HIV Practice Course: Building Capacity of Healthcare Workers in the Region

BackgroundBuilding the capacity of healthcare workers (HCW) can positively influence service quality and patient care. Given the limited HIV training opportunities in the Asia Pacific Region, the Asia Pacific HIV Practice Course (APHPC) aims to improve knowledge and skills and encourage patient-centered practice.MethodsThe APHPC is organized by an interprofessional organizing committee. The course was developed based on a needs assessment of HCWs in the region and is run over 4 days. Using didactic and interactive learning approaches, the course covers psychosocial, biomedical and service delivery topics. One hundred HCWs, predominantly nurses, social workers/counselors and, pharmacists have attended the course over 2 years (2016 and 2017). Participants have attended from Singapore, Sri Lanka, Indonesia, Malaysia, Philippines, Bangladesh, Papua New Guinea, Myanmar, and Vietnam. Pre- and post-course evaluations were completed, as well as daily session evaluations and a 3-month post-course evaluation. Participants rated their confidence from 1 to 5 (not confident–very confident). A paired-samples t-test was conducted to compare mean (M) confidence levels pre- and post-course in the various topics covered.ResultsThere was significant improvement in mean confidence scores pre-course (M = 3.40, SD 0.27) and post-course (M = 4.09, SD = 0.13); t(11)= 13.1958, P < 0.0001. This was seen across all topics (figure) with the most marked improvement found in models of care and barriers to care, topics that are not routinely or explicitly covered in the same detail or frequency as HIV basics, testing, treatment and prevention. In the 3-month post-course survey, participants shared the changes they have made to their local practice since attending the APHPC. Changes include, additional staff training, revisions to counseling models and services and the implementation of monitoring and evaluation structures.ConclusionThe APHPC has proven to be an impactful and highly evaluated course. To ensure the course continues to influence and improve practice, the content of the course can be expanded to cover nonstandard topics, and further interactive learning experiences can be incorporated.Disclosures All authors: No reported disclosures.

Students’ Collaborative Learning Attitudes and Their Satisfaction with Online Collaborative Case-Based Courses

ABSTRACTCase-based learning is an interactive group learning approach that embodies the principles of Socratic inquiry. It is a common pedagogy used in many disciplines including business, law, and health. With online learning being the fastest growing segment of higher education, case-based learning is also finding its way into online courses. This research was conducted on a population of 698 students enrolled in online case-based courses. The use of Structural Equation Modelling provided an understanding of the factors that influence student satisfaction. Through Logistics Regression, the extent to which these factors do influence satisfaction was determined. The findings indicate that the greatest predictors of student satisfaction were the interaction and understanding between the students and the instructor. Furthermore, a student’s pre-existing attitude toward interactive group learning is also a significant factor in predicting student satisfaction. The conclusions point to a lack of understanding by students on how to engage in effective interactive online group learning. The recommendation is that training in online collaborative skills be considered as a front-end to any online case-based course.

Education and Training: Key Factors in Global Occupational and Environmental Health.

Introduction:Education and training in Occupational and Environmental Health (OEH) play an important role in building global capacity and contribute to safer working conditions. The shortage of occupational health professionals, the lack of knowledge and a high number of occupational accidents and diseases stress the demand for providing further education and training in OEH. This need is especially urgent in low and middle-income countries. Three international courses on OEH provide valuable insights on how to develop successful trainings in the field and how to contribute to the creation of healthy and safe workplaces.Methodology and Results:The courses “Global Occupational and Environmental Determinants of Diseases: a multidisciplinary and multicultural approach for prevention” (University of Brescia, Italy), “Advanced International Training Course in OEH” (Chulabhorn Research Institute, Bangkok, Thailand) and “Teaching interventions crossing borders” (Ludwig-Maximillians-Universität, Munich, Germany) offer a unique networking opportunity for health professionals from low-, middle- and high-income countries. Three main characteristics of learning were identified as successful for adult learning in OEH: Interdisciplinary learning allows exchanging skills and knowledge and therefore closes gaps between the several disciplines within OEH. Experiential learning enables participants to reflect on their personal experiences, e.g. through workplace visits. Finally, active learning empowers the learner to take the leading role in the learning process using e.g. problem- or project-based learning approaches.Conclusion:Training and education in OEH should become a higher priority by including it in the standard public health and medical curricula. For this, train the trainer approaches are essential in many countries. Any OEH training should follow the interdisciplinary, experiential and interactive learning approach and should be accessible to participants from all regions.

Combination of learning from non-optimal demonstrations and feedbacks using inverse reinforcement learning and Bayesian policy improvement

Inverse reinforcement learning (IRL) is a powerful tool for teaching by demonstrations, provided that sufficiently diverse and optimal demonstrations are given, and learner agent correctly perceives those demonstrations. These conditions are hard to meet in practice; as a trainer cannot cover all possibilities by demonstrations, he may partially fail to follow the optimal behavior. Also, trainer and learner have different perceptions of the environment including trainer's actions. A practical way to overcome these problems is using a combination of trainer's demonstrations and feedbacks.We propose an interactive learning approach to overcome the challenge of non-optimal demonstrations by integrating human evaluative feedbacks with the IRL process, given sufficiently diverse demonstrations and the domain transition model. To this end, we develop a probabilistic model of human feedbacks and iteratively improve the agent policy using Bayes rule. We then integrate this information in an extended IRL algorithm to enhance the learned reward function.We examine the developed approach in one experimental and two simulated tasks; i.e., a grid world navigation, a highway car driving system and a navigation task by the e-puck robot. Obtained results show significant improved efficiency of the proposed approach in face of having different levels of non-optimality in demonstrations and the number of evaluative feedbacks.

Motion-Sound Mapping through Interaction

Technologies for sensing movement are expanding toward everyday use in virtual reality, gaming, and artistic practices. In this context, there is a need for methodologies to help designers and users create meaningful movement experiences. This article discusses a user-centered approach for the design of interactive auditory feedback using interactive machine learning. We discuss Mapping through Interaction, a method for crafting sonic interactions from corporeal demonstrations of embodied associations between motion and sound. It uses an interactive machine learning approach to build the mapping from user demonstrations, emphasizing an iterative design process that integrates acted and interactive experiences of the relationships between movement and sound. We examine Gaussian Mixture Regression and Hidden Markov Regression for continuous movement recognition and real-time sound parameter generation. We illustrate and evaluate this approach through an application in which novice users can create interactive sound feedback based on coproduced gestures and vocalizations. Results indicate that Gaussian Mixture Regression and Hidden Markov Regression can efficiently learn complex motion-sound mappings from few examples.

Effects of interactive teaching on university students' knowledge and attitude toward reproductive health: a pilot study in Jordan.

BackgroundYouths in Jordan lack knowledge related to reproductive health (RH). Interactive teaching methods showed positive results in enhancing health awareness and adopting healthy practices among students.ObjectivesThe objective of this study was to examine the usefulness of interactive teaching in promoting health awareness of RH among nonmedical university students in Jordan.MethodsWe employed a quasi-experimental one group pretest and posttest design for a purposive sample of 210 students (18–24 years). Knowledge and attitudes regarding RH issues were assessed using a questionnaire developed by the researchers.ResultsA significant improvement in students’ knowledge and attitudes toward RH was evident. Female students had higher scores on knowledge than male students in the pretest; this difference was smaller in the posttest. Also, female students had significantly more positive attitudes toward RH in pretest than males, although this difference vanished in the posttest. Study results indicated that students benefit from study intervention regardless their gender.ConclusionIntegrating RH into university’s curriculum coupled with interactive learning approach is a powerful way to promote RH awareness among youths.

A Student‐Centered, Active Learning Approach to Teaching Spinal Cord Anatomy

PurposeThe purpose of this project was to employ and preliminarily evaluate the effectiveness of small group, active learning among second‐year medical students studying the anatomy of the spinal cord.MethodsWe presented spinal cord anatomy in our “Human Condition” course for two cohorts of second‐year medical students. In 2015, and years prior, the material was incorporated into large group sessions and presented using slides in a large lecture hall. In 2017, we reorganized this learning activity into a more student‐centered, small group laboratory session. The interactive learning approach emphasized group work and higher‐order thinking to apply knowledge gained from assigned readings. In addition, students were encouraged to correlate anatomical features of prosected human spinal cord specimens with histological sections. After working together to solve the prompts posed by the faculty member, the students applied the teach‐back method to explain their answers and rationales to their facilitator. Facilitators used Socratic questioning to guide the discussions and to promote higher‐order thinking. Course examinations from both years included one question about spinal cord anatomy relevant to the material covered in both sessions. We compared scores from both years to evaluate the effectiveness of these learning approaches.ResultsStudents who participated in the student‐centered laboratory approach demonstrated a modest increase (3%) in overall performance on the spinal cord examination question. The transition was faculty‐intensive, requiring both an increase in the number of trained facilitators and multiple iterations of the same station.ConclusionsThe student‐centered, active learning approach was associated with a modest increase in student performance on the course examination. From our preliminary evaluation, we observed no detrimental effects from shifting the large group session to small group lab stations; indeed, our evaluation suggests at least a modest benefit. The use of spinal cord specimens underscores the importance of donated anatomical material for medical education. These specimens were integral to the learning activity, as students applied their observations of gross features to their decision‐making in solving prompts relevant to spinal cord histology. Additionally, the small group laboratory approach was an efficient use of curricular time. We aim to continue to study the effectiveness of this learning approach, ideally migrating other large group sessions into the laboratory whenever possible.This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.

Blackthorn: Large-Scale Interactive Multimodal Learning

This paper presents Blackthorn, an efficient interactive multimodal learning approach facilitating analysis of multimedia collections of up to 100 million items on a single high-end workstation. Blackthorn features efficient data compression, feature selection, and optimizations to the interactive learning process. The Ratio-64 data representation introduced in this paper only costs tens of bytes per item yet preserves most of the visual and textual semantic information with good accuracy. The optimized interactive learning model scores the Ratio-64-compressed data directly, greatly reducing the computational requirements. The experiments compare Blackthorn with two baselines: Conventional relevance feedback, and relevance feedback using product quantization to compress the features. The results show that Blackthorn is up to 77.5 $\times$ faster than the conventional relevance feedback alternative, while outperforming the baseline with respect to the relevance of results: It vastly outperforms the baseline on recall over time and reaches up to 108% of its precision. Compared to the product quantization variant, Blackthorn is just as fast, while producing more relevant results. On the full YFCC100M dataset, Blackthorn performs one complete interaction round in roughly 1 s while maintaining adequate relevance of results, thus opening multimedia collections comprising up to 100 million items to fully interactive learning-based analysis.

An interactive and blended learning model for engineering education

Higher education has been undergoing a global transformation since the 1990s, resulting in changes, induced and modulated by several factors. The most important of those changes include the significant technological development of the Internet and handheld devices, along with the growing demand for educational program assessments by accreditation bodies and funding/government agencies. The involvement of third-party resources (e.g., book publishers) and technology (e.g., interactive SMART Board) have induced further positive influences within higher education and advanced the learning experience to the next level. These factors and others have inspired many innovative teaching proposals to improve the learning experience and to engage students in class more effectively. This article proposes an interactive learning approach for teaching engineering curriculum. The proposed teaching model takes advantage of the blended learning approach in the form of flipped classrooms to complement the learning experience in engineering design courses. The flipped classroom is focused on moving some components of the traditional face-to-face lecture out of class. With a flipped classroom, students can prepare for class by interacting with provided resources to reduce lecture time and allow more engagement in class. This work also demonstrates how to apply constructivism and conversation theories to enhance learning in a blended learning environment. The teaching method is carried out by developing an interactive module for each lecture and a mini-lecture video, which is provided to review materials before class or as a lecture recording. Each lecture module consists of pre-class brainstorming questions, a lecture theory, and of practice exercises, which are arranged in a series of guided instructional solution steps. The lecture module will be delivered to students using interactive presentation software that engages students in class activities. The effectiveness of this teaching method is demonstrated through a quantitative assessment of program learning outcomes (PLO) before and after applying the method. Assessment of learning outcomes indicates an average improvement from 3.9 to 4.4 on a scale of 1–5.

Communication Issues in End-of-Life Care - An African Setting Experience

Communications with patients and family around End-of-life care issues in any society could be very challenging. Hitherto, in some African settings communicating bad or death news is usually handled and effectively managed by the support of the community. Clinical bad news, dying and death in the hospital setting can only be best approached in an interdisciplinary or by a multidisciplinary team. Bad news is a message, with a potential of shattering hopes and dreams that often times provoke cognitive, behavioural, affective and spiritual distress and pains as evident in some life-threatening illnesses like Cancer and HIV/AIDS. Clinical competencies should include skills to break bad or difficult news tactically to ensure better psychological adjustment by the patient and family members, and also reduces stress on everyone including the health care professionals. The clinicians can make the news worse if not empathically delivered and also have the task of empowering the patient towards attaining a realistic hope. Adjustment to bad news may take time and is similar to a process of grief or loss. Bad or difficult news communication training to health care providers has largely been limited in Nigeria. An interactive experiential adult learning approach can be effective in communication skills training for health care providers especially among palliative care practitioners as most clinicians missed this skill at the undergraduate training.

Improving the Working Memory During Early Childhood Education Through the Use of an Interactive Gesture Game-Based Learning Approach

One of the most socially and culturally advantageous uses of human-computer interaction is enhancing playing and learning for children. In this paper, gesture interactive game-based learning (GIGL) is tested to see if these kinds of applications are suitable to stimulate working memory (WM) and basic mathematical skills (BMS) in early childhood (5-6 years old) using a hand gesture recognition system. Hand gesture is being performed by the user and to control a computer system by that incoming information. The research was developed using a quasi-experimental design with a pre-test and post-test, using both an experimental and control group through three phases: the first one was the prior evaluation of the learner's skills; a second phase in which the use of the technology was developed; and a final phase of evaluation. In the evaluation phases, working memory was measured using the Corsi task, and the basic mathematical skills using the test for the diagnosis of basic mathematical competencies (TEDI-MATH). The results provide clear evidence that the use of these technologies improved both working memory and basic mathematical skills. We can conclude that the children who used GIGL technology showed a significant increase in their learning performance in WM and BMS, surpassing those who did normal school activities.

Abjad: Towards Interactive Learning Approach to Arabic Reading Based on Speech Recognition

Speech recognition technology has gained more attention recently especially with the increasing use of mobile phones, and tablets. Yet, researches in speech recognition technology for Arabic language is still limited. This paper presents our experience with Arabic speech recognition using Android’s SpeechRecognizer Application Program Interface (API) through the design and implementation of Abjad.Abjad is a mobile application for learning Arabic reading which mainly targets children at the age of six years old. It interacts with the users in a real-time manner by prompting them to record their voice while reading. The voice is then instantly evaluated and vocal feedback is provided according the accuracy of the user reading. In addition, the users can do tests to assess their progress in reading. The main goals of Abjad are to provide the Arabic language learner in general, and primary one children in particular with interactive technology to shorten their learning journey. More importantly, to enrich the Arabic speech recognition researches with our experience using SpeechRecpgnizer API for Arabic language, including the optimization techniques we developed to increase the accuracy rate of Arabic speech recognition from 44.44% to up to 94.44% with adults and 80% with children.

Interactive online machine learning approach for activity-travel survey

This article proposes a framework for an interactive activity-travel survey method, implementable on mobile devices such as smartphones. The proposed method was developed to reduce the burden (i.e., frequency of questions) on respondents in long-term behavioral surveys, without relying on external data sources. The method employs an online travel context estimation model and an online machine learning method as interactive processes. The estimation model is used for automatically estimating travel contexts during surveys, while the online machine learning method is used for dynamically updating the estimation model, based on answers from respondents. The proposed method was examined by simulations using data obtained from a past probe person survey. The results suggest that the frequency of inputs by respondents in surveys can be significantly reduced, while maintaining high accuracy of the obtained data. For example, the method successfully estimated certain types of trips (e.g., commuting) and the behaviors of certain respondents (e.g., those whose activity-travel pattern is recurrent) because of the learning process and reduced survey burden on them. Meanwhile, although the method could not always precisely estimate some other types of trips, it eventually obtained accurate results because of the interaction process. Therefore, the proposed method could be useful to reduce the burden on respondents in long-term surveys, while maintaining high data quality and capturing traveler heterogeneity.

Relation Between Interactive Learning and Prior Knowledge: Insights from a General Education Program of Science and Humanities

Abstract Possession of prior knowledge has been shown to critically influence students' learning in major courses. Yet the topic remains unexplored in the realm of general education in higher education. The study aims at investigating whether students lacking relevant high school knowledge background could benefit in the two courses, one inclined to science and one inclined to the humanities, in a general education program at a university in Hong Kong. Interactive learning has been intensively practiced in the program to minimize the unfavorable impact brought up by a lack of prior knowledge among students. Data from an entry-exit survey show that a lack of prior knowledge does not limit students' learning in outcomes related to generic cognitive skills and subject matter content in both courses. An interactive learning approach is suggested to have mediated the problem of the lack of prior knowledge in students' subsequent learning in the two general education courses.

Diversity-based interactive learning meets multimodality

In interactive retrieval tasks, one of the main objectives is to maximize the user information gain throughout search sessions. Retrieving many relevant items is quite important, but it does not necessarily completely satisfy the user needs. When only relevant near-duplicate items are retrieved, the amount of different concepts users are able to extract from the target collection is very limited. Therefore, broadening the number of concepts present in a result set may improve the overall search experience. Diversifying concepts present in the retrieved set is one possibility for increasing the information gain in a single search iteration, maximizing the likelihood of including at least some relevant items for each possible intent of ambiguous or underspecified queries. Relevance feedback approaches may also take advantage of diverse results to improve internal machine learning models. In this context, this work proposes and analyses several multimodal image retrieval approaches built over a learning framework for relevance feedback on diversified results. Our experimental analysis shows that different retrieval modalities are positively impacted by diversity, but achieve best retrieval effectiveness with diversification applied at different moments of a search session. Moreover, the best results are achieved with a query-by-example approach using multimodal information obtained from feedback. In summary, we demonstrate that learning with diversity is an effective alternative for boosting multimodal interactive learning approaches.