Palakorn Achananuparp

Shan Dong, Palakorn Achananuparp, Hieu Hien Mai et al.

In this work, we develop a novel reasoning approach to enhance the performance of large language models (LLMs) in future occupation prediction. In this approach, a reason generator first derives a ``reason'' for a user using his/her past education and career history. The reason summarizes the user's preference and is used as the input of an occupation predictor to recommend the user's next occupation. This two-step occupation prediction approach is, however, non-trivial as LLMs are not aligned with career paths or the unobserved reasons behind each occupation decision. We therefore propose to fine-tune LLMs improving their reasoning and occupation prediction performance. We first derive high-quality oracle reasons, as measured by factuality, coherence and utility criteria, using a LLM-as-a-Judge. These oracle reasons are then used to fine-tune small LLMs to perform reason generation and next occupation prediction. Our extensive experiments show that: (a) our approach effectively enhances LLM's accuracy in next occupation prediction making them comparable to fully supervised methods and outperforming unsupervised methods; (b) a single LLM fine-tuned to perform reason generation and occupation prediction outperforms two LLMs fine-tuned to perform the tasks separately; and (c) the next occupation prediction accuracy depends on the quality of generated reasons. Our code is available at https://github.com/Sarasarahhhhh/job_prediction.

Suhyun Lee, Palakorn Achananuparp, Neemesh Yadav et al.

Large language models (LLMs) are increasingly explored as scalable tools for mental health counseling, yet evaluating their safety remains challenging due to the interactional and context-dependent nature of clinical harm. Existing evaluation frameworks predominantly assess isolated responses using coarse-grained taxonomies or static datasets, limiting their ability to diagnose how harms emerge and accumulate over multi-turn counseling interactions. In this work, we introduce R-MHSafe, a role-aware mental health safety taxonomy that characterizes clinically significant harm in terms of the interactional roles an AI counselor adopts, including perpetrator, instigator, facilitator, or enabler, combined with clinically grounded harm categories. Then, we propose MHSafeEval, a closed-loop, agent-based evaluation framework that formulates safety assessment as trajectory-level discovery of harm through adversarial multi-turn interactions, guided by role-aware modeling. Using R-MHSafe and MHSafeEval, we conduct a large-scale evaluation across state-of-the-art LLMs. Our results reveal substantial role-dependent and cumulative safety failures that are systematically missed by existing static benchmarks, and show that our framework significantly improves failure-mode coverage and diagnostic granularity.

Palakorn Achananuparp, Connie Xu, Yao Lu et al.

We present a large-scale analysis of career mobility of college-educated U.S. workers using online resume profiles to investigate how gender, race, and job change options are associated with upward mobility. This study addresses key research questions of how the job changes affect their upward career mobility, and how the outcomes of upward career mobility differ by gender and race. We address data challenges -- such as missing demographic attributes, missing wage data, and noisy occupation labels -- through various data processing and Artificial Intelligence (AI) methods. In particular, we develop a large language models (LLMs) based occupation classification method known as FewSOC that achieves accuracy significantly higher than the original occupation labels in the resume dataset. Analysis of 228,710 career trajectories reveals that intra-firm occupation change has been found to facilitate upward mobility most strongly, followed by inter-firm occupation change and inter-firm lateral move. Women and Black college graduates experience significantly lower returns from job changes than men and White peers. Multilevel sensitivity analyses confirm that these disparities are robust to cluster-level heterogeneity and reveal additional intersectional patterns.

Neemesh Yadav, Palakorn Achananuparp, Jing Jiang et al.

Large Language Models (LLMs) have been shown to possess Theory of Mind (ToM) abilities. However, it remains unclear whether this stems from robust reasoning or spurious correlations. We introduce DialToM, a human-verified benchmark built from natural human dialogue using a multiple-choice framework. We evaluate not only mental state prediction (Literal ToM) but also the functional utility of these states (Functional ToM) through Prospective Diagnostic Forecasting -- probing whether models can identify state-consistent dialogue trajectories solely from mental-state profiles. Our results reveal a significant reasoning asymmetry: while LLMs excel at identifying mental states, most (except for Gemini 3 Pro) fail to leverage this understanding to forecast social trajectories. Additionally, we find only weak semantic similarities between human and LLM-generated inferences. To facilitate reproducibility, the DialToM dataset and evaluation code are publicly available at https://github.com/Stealth-py/DialToM.

Neemesh Yadav, Palakorn Achananuparp, Jing Jiang et al.

Large Language Models (LLMs) have shown potential in simulating human behaviors and performing theory-of-mind (ToM) reasoning, a crucial skill for complex social interactions. In this study, we investigate the role of ToM reasoning in aligning agentic behaviors with human norms in negotiation tasks, using the ultimatum game as a controlled environment. We initialized LLM agents with different prosocial beliefs (including Greedy, Fair, and Selfless) and reasoning methods like chain-of-thought (CoT) and varying ToM levels, and examined their decision-making processes across diverse LLMs, including reasoning models like o3-mini and DeepSeek-R1 Distilled Qwen 32B. Results from 2,700 simulations indicated that ToM reasoning enhances behavior alignment, decision-making consistency, and negotiation outcomes. Consistent with previous findings, reasoning models exhibit limited capability compared to models with ToM reasoning, different roles of the game benefits with different orders of ToM reasoning. Our findings contribute to the understanding of ToM's role in enhancing human-AI interaction and cooperative decision-making. The code used for our experiments can be found at https://github.com/Stealth-py/UltimatumToM.

Philips Kokoh Prasetyo, Palakorn Achananuparp, Ee-Peng Lim

Recent research has identified a few design flaws in popular mobile health (mHealth) applications for promoting healthy eating lifestyle, such as mobile food journals. These include tediousness of manual food logging, inadequate food database coverage, and a lack of healthy dietary goal setting. To address these issues, we present Foodbot, a chatbot-based mHealth application for goal-oriented just-in-time (JIT) healthy eating interventions. Powered by a large-scale food knowledge graph, Foodbot utilizes automatic speech recognition and mobile messaging interface to record food intake. Moreover, Foodbot allows users to set goals and guides their behavior toward the goals via JIT notification prompts, interactive dialogues, and personalized recommendation. Altogether, the Foodbot framework demonstrates the use of open-source data, tools, and platforms to build a practical mHealth solution for supporting healthy eating lifestyle in the general population.

Yizhe Yang, Palakorn Achananuparp, Heyan Huang et al.

The recent success of large language models (LLMs) has attracted widespread interest to develop role-playing conversational agents personalized to the characteristics and styles of different speakers to enhance their abilities to perform both general and special purpose dialogue tasks. However, the ability to personalize the generated utterances to speakers, whether conducted by human or LLM, has not been well studied. To bridge this gap, our study introduces a novel evaluation challenge: speaker verification in agent-generated conversations, which aimed to verify whether two sets of utterances originate from the same speaker. To this end, we assemble a large dataset collection encompassing thousands of speakers and their utterances. We also develop and evaluate speaker verification models under experiment setups. We further utilize the speaker verification models to evaluate the personalization abilities of LLM-based role-playing models. Comprehensive experiments suggest that the current role-playing models fail in accurately mimicking speakers, primarily due to their inherent linguistic characteristics.

Yizhe Yang, Palakorn Achananuparp, Heyan Huang et al.

Simulating human clients in mental health counseling is crucial for training and evaluating counselors (both human or simulated) in a scalable manner. Nevertheless, past research on client simulation did not focus on complex conversation tasks such as mental health counseling. In these tasks, the challenge is to ensure that the client's actions (i.e., interactions with the counselor) are consistent with with its stipulated profiles and negative behavior settings. In this paper, we propose a novel framework that supports consistent client simulation for mental health counseling. Our framework tracks the mental state of a simulated client, controls its state transitions, and generates for each state behaviors consistent with the client's motivation, beliefs, preferred plan to change, and receptivity. By varying the client profile and receptivity, we demonstrate that consistent simulated clients for different counseling scenarios can be effectively created. Both our automatic and expert evaluations on the generated counseling sessions also show that our client simulation method achieves higher consistency than previous methods.

Yizhe Yang, Palakorn Achananuparp, Heyan Huang et al.

Conversational counselor agents have become essential tools for addressing the rising demand for scalable and accessible mental health support. This paper introduces CAMI, a novel automated counselor agent grounded in Motivational Interviewing (MI) -- a client-centered counseling approach designed to address ambivalence and facilitate behavior change. CAMI employs a novel STAR framework, consisting of client's state inference, motivation topic exploration, and response generation modules, leveraging large language models (LLMs). These components work together to evoke change talk, aligning with MI principles and improving counseling outcomes for clients from diverse backgrounds. We evaluate CAMI's performance through both automated and manual evaluations, utilizing simulated clients to assess MI skill competency, client's state inference accuracy, topic exploration proficiency, and overall counseling success. Results show that CAMI not only outperforms several state-of-the-art methods but also shows more realistic counselor-like behavior. Additionally, our ablation study underscores the critical roles of state inference and topic exploration in achieving this performance.

Palakorn Achananuparp, Ee-Peng Lim, Yao Lu

Automatically annotating job data with standardized occupations from taxonomies, known as occupation classification, is crucial for labor market analysis. However, this task is often hindered by data scarcity and the challenges of manual annotations. While large language models (LLMs) hold promise due to their extensive world knowledge and in-context learning capabilities, their effectiveness depends on their knowledge of occupational taxonomies, which remains unclear. In this study, we assess the ability of LLMs to generate precise taxonomic entities from taxonomy, highlighting their limitations, especially for smaller models. To address these challenges, we propose a multi-stage framework consisting of inference, retrieval, and reranking stages, which integrates taxonomy-guided reasoning examples to enhance performance by aligning outputs with taxonomic knowledge. Evaluations on a large-scale dataset show that our framework not only enhances occupation and skill classification tasks, but also provides a cost-effective alternative to frontier models like GPT-4o, significantly reducing computational costs while maintaining strong performance. This makes it a practical and scalable solution for occupation classification and related tasks across LLMs.

V N S Rama Krishna Pinnimty, Matt Zhao, Palakorn Achananuparp et al.

We present an invert-and-edit framework to automatically transform facial weight of an input face image to look thinner or heavier by leveraging semantic facial attributes encoded in the latent space of Generative Adversarial Networks (GANs). Using a pre-trained StyleGAN as the underlying generator, we first employ an optimization-based embedding method to invert the input image into the StyleGAN latent space. Then, we identify the facial-weight attribute direction in the latent space via supervised learning and edit the inverted latent code by moving it positively or negatively along the extracted feature axis. Our framework is empirically shown to produce high-quality and realistic facial-weight transformations without requiring training GANs with a large amount of labeled face images from scratch. Ultimately, our framework can be utilized as part of an intervention to motivate individuals to make healthier food choices by visualizing the future impacts of their behavior on appearance.

Hao Wang, Doyen Sahoo, Chenghao Liu et al.

Food retrieval is an important task to perform analysis of food-related information, where we are interested in retrieving relevant information about the queried food item such as ingredients, cooking instructions, etc. In this paper, we investigate cross-modal retrieval between food images and cooking recipes. The goal is to learn an embedding of images and recipes in a common feature space, such that the corresponding image-recipe embeddings lie close to one another. Two major challenges in addressing this problem are 1) large intra-variance and small inter-variance across cross-modal food data; and 2) difficulties in obtaining discriminative recipe representations. To address these two problems, we propose Semantic-Consistent and Attention-based Networks (SCAN), which regularize the embeddings of the two modalities through aligning output semantic probabilities. Besides, we exploit a self-attention mechanism to improve the embedding of recipes. We evaluate the performance of the proposed method on the large-scale Recipe1M dataset, and show that we can outperform several state-of-the-art cross-modal retrieval strategies for food images and cooking recipes by a significant margin.

Helena H. Lee, Ke Shu, Palakorn Achananuparp et al.

Interests in the automatic generation of cooking recipes have been growing steadily over the past few years thanks to a large amount of online cooking recipes. We present RecipeGPT, a novel online recipe generation and evaluation system. The system provides two modes of text generations: (1) instruction generation from given recipe title and ingredients; and (2) ingredient generation from recipe title and cooking instructions. Its back-end text generation module comprises a generative pre-trained language model GPT-2 fine-tuned on a large cooking recipe dataset. Moreover, the recipe evaluation module allows the users to conveniently inspect the quality of the generated recipe contents and store the results for future reference. RecipeGPT can be accessed online at https://recipegpt.org/.

Doyen Sahoo, Wang Hao, Shu Ke et al.

An important aspect of health monitoring is effective logging of food consumption. This can help management of diet-related diseases like obesity, diabetes, and even cardiovascular diseases. Moreover, food logging can help fitness enthusiasts, and people who wanting to achieve a target weight. However, food-logging is cumbersome, and requires not only taking additional effort to note down the food item consumed regularly, but also sufficient knowledge of the food item consumed (which is difficult due to the availability of a wide variety of cuisines). With increasing reliance on smart devices, we exploit the convenience offered through the use of smart phones and propose a smart-food logging system: FoodAI, which offers state-of-the-art deep-learning based image recognition capabilities. FoodAI has been developed in Singapore and is particularly focused on food items commonly consumed in Singapore. FoodAI models were trained on a corpus of 400,000 food images from 756 different classes. In this paper we present extensive analysis and insights into the development of this system. FoodAI has been deployed as an API service and is one of the components powering Healthy 365, a mobile app developed by Singapore's Heath Promotion Board. We have over 100 registered organizations (universities, companies, start-ups) subscribing to this service and actively receive several API requests a day. FoodAI has made food logging convenient, aiding smart consumption and a healthy lifestyle.

Helena Lee, Palakorn Achananuparp, Yue Liu et al.

Consumption of diets with low glycemic impact is highly recommended for diabetics and pre-diabetics as it helps maintain their blood glucose levels. However, laboratory analysis of dietary glycemic potency is time-consuming and expensive. In this paper, we explore a data-driven approach utilizing online crowdsourcing and machine learning to estimate the glycemic impact of cooking recipes. We show that a commonly used healthiness metric may not always be effective in determining recipes suitable for diabetics, thus emphasizing the importance of the glycemic-impact estimation task. Our best classification model, trained on nutritional and crowdsourced data obtained from Amazon Mechanical Turk (AMT), can accurately identify recipes which are unhealthful for diabetics.

Yue Liu, Helena Lee, Palakorn Achananuparp et al.

Human beings are creatures of habit. In their daily life, people tend to repeatedly consume similar types of food items over several days and occasionally switch to consuming different types of items when the consumptions become overly monotonous. However, the novel and repeat consumption behaviors have not been studied in food recommendation research. More importantly, the ability to predict daily eating habits of individuals is crucial to improve the effectiveness of food recommender systems in facilitating healthy lifestyle change. In this study, we analyze the patterns of repeat food consumptions using large-scale consumption data from a popular online fitness community called MyFitnessPal (MFP), conduct an offline evaluation of various state-of-the-art algorithms in predicting the next-day food consumption, and analyze their performance across different demographic groups and contexts. The experiment results show that algorithms incorporating the exploration-and-exploitation and temporal dynamics are more effective in the next-day recommendation task than most state-of-the-art algorithms.

Palakorn Achananuparp, Ee-Peng Lim, Vibhanshu Abhishek

Past research has shown the benefits of food journaling in promoting mindful eating and healthier food choices. However, the links between journaling and healthy eating have not been thoroughly examined. Beyond caloric restriction, do journalers consistently and sufficiently consume healthful diets? How different are their eating habits compared to those of average consumers who tend to be less conscious about health? In this study, we analyze the healthy eating behaviors of active food journalers using data from MyFitnessPal. Surprisingly, our findings show that food journalers do not eat as healthily as they should despite their proclivity to health eating and their food choices resemble those of the general populace. Furthermore, we find that the journaling duration is only a marginal determinant of healthy eating outcomes and sociodemographic factors, such as gender and regions of residence, are much more predictive of healthy food choices.

Palakorn Achananuparp, Ingmar Weber

In this paper, we explore the problem of identifying substitute relationship between food pairs from real-world food consumption data as the first step towards the healthier food recommendation. Our method is inspired by the distributional hypothesis in linguistics. Specifically, we assume that foods that are consumed in similar contexts are more likely to be similar dietarily. For example, a turkey sandwich can be considered a suitable substitute for a chicken sandwich if both tend to be consumed with french fries and salad. To evaluate our method, we constructed a real-world food consumption dataset from MyFitnessPal's public food diary entries and obtained ground-truth human judgements of food substitutes from a crowdsourcing service. The experiment results suggest the effectiveness of the method in identifying suitable substitutes.

Ingmar Weber, Palakorn Achananuparp

To support people trying to lose weight and stay healthy, more and more fitness apps have sprung up including the ability to track both calories intake and expenditure. Users of such apps are part of a wider ``quantified self'' movement and many opt-in to publicly share their logged data. In this paper, we use public food diaries of more than 4,000 long-term active MyFitnessPal users to study the characteristics of a (un-)successful diet. Concretely, we train a machine learning model to predict repeatedly being over or under self-set daily calories goals and then look at which features contribute to the model's prediction. Our findings include both expected results, such as the token ``mcdonalds'' or the category ``dessert'' being indicative for being over the calories goal, but also less obvious ones such as the difference between pork and poultry concerning dieting success, or the use of the ``quick added calories'' functionality being indicative of over-shooting calorie-wise. This study also hints at the feasibility of using such data for more in-depth data mining, e.g., looking at the interaction between consumed foods such as mixing protein- and carbohydrate-rich foods. To the best of our knowledge, this is the first systematic study of public food diaries.