Search

We introduce Phi-4-Mini and Phi-4-Multimodal, compact yet highly capable language and multimodal models. Phi-4-Mini is a 3.8-billion-parameter language model trained on high-quality web and synthetic data, significantly outperforming recent open-source models of similar size and matching the performance of models twice its size on math and coding tasks requiring complex reasoning. This achievement is driven by a carefully curated synthetic data recipe emphasizing high-quality math and coding datasets. Compared to its predecessor, Phi-3.5-Mini, Phi-4-Mini features an expanded vocabulary size of 200K tokens to better support multilingual applications, as well as group query attention for more efficient long-sequence generation. Phi-4-Multimodal is a multimodal model that integrates text, vision, and speech/audio input modalities into a single model. Its novel modality extension approach leverages LoRA adapters and modality-specific routers to allow multiple inference modes combining various modalities without interference. For example, it now ranks first in the OpenASR leaderboard to date, although the LoRA component of the speech/audio modality has just 460 million parameters. Phi-4-Multimodal supports scenarios involving (vision + language), (vision + speech), and (speech/audio) inputs, outperforming larger vision-language and speech-language models on a wide range of tasks. Additionally, we experiment to further train Phi-4-Mini to enhance its reasoning capabilities. Despite its compact 3.8-billion-parameter size, this experimental version achieves reasoning performance on par with or surpassing significantly larger models, including DeepSeek-R1-Distill-Qwen-7B and DeepSeek-R1-Distill-Llama-8B., Comment: 39 pages

Large Language Models (LLMs) have made notable progress in mathematical reasoning, yet they often rely on single-paradigm reasoning that limits their effectiveness across diverse tasks. In this paper, we introduce Chain-of-Reasoning (CoR), a novel unified framework that integrates multiple reasoning paradigms--Natural Language Reasoning (NLR), Algorithmic Reasoning (AR), and Symbolic Reasoning (SR)--to enable synergistic collaboration. CoR generates multiple potential answers using different reasoning paradigms and synthesizes them into a coherent final solution. We propose a Progressive Paradigm Training (PPT) strategy that allows models to progressively master these paradigms, culminating in the development of CoR-Math-7B. Experimental results demonstrate that CoR-Math-7B significantly outperforms current SOTA models, achieving up to a 41.0% absolute improvement over GPT-4 in theorem proving tasks and a 7.9% improvement over RL-based methods in arithmetic tasks. These results showcase the enhanced mathematical comprehensive ability of our model, achieving significant performance gains on specific tasks and enabling zero-shot generalization across tasks.

The scaling laws and extraordinary performance of large foundation models motivate the development and utilization of such models in biomedicine. However, despite early promising results on some biomedical benchmarks, there are still major challenges that need to be addressed before these models can be used in real-world clinics. Frontier general-domain models such as GPT-4V still have significant performance gaps in multimodal biomedical applications. More importantly, less-acknowledged pragmatic issues, including accessibility, model cost, and tedious manual evaluation make it hard for clinicians to use state-of-the-art large models directly on private patient data. Here, we explore training open-source small multimodal models (SMMs) to bridge competency gaps for unmet clinical needs in radiology. To maximize data efficiency, we adopt a modular approach by incorporating state-of-the-art pre-trained models for image and text modalities, and focusing on training a lightweight adapter to ground each modality to the text embedding space, as exemplified by LLaVA-Med. For training, we assemble a large dataset of over 697 thousand radiology image-text pairs. For evaluation, we propose CheXprompt, a GPT-4-based metric for factuality evaluation, and demonstrate its parity with expert evaluation. For best practice, we conduct a systematic ablation study on various choices in data engineering and multimodal training. The resulting LlaVA-Rad (7B) model attains state-of-the-art results on standard radiology tasks such as report generation and cross-modal retrieval, even outperforming much larger models such as GPT-4V and Med-PaLM M (84B). The inference of LlaVA-Rad is fast and can be performed on a single V100 GPU in private settings, offering a promising state-of-the-art tool for real-world clinical applications.

Artificial General Intelligence (AGI) requires comprehensive understanding and generation capabilities for a variety of tasks spanning different modalities and functionalities. Integrative AI is one important direction to approach AGI, through combining multiple models to tackle complex multimodal tasks. However, there is a lack of a flexible and composable platform to facilitate efficient and effective model composition and coordination. In this paper, we propose the i-Code Studio, a configurable and composable framework for Integrative AI. The i-Code Studio orchestrates multiple pre-trained models in a finetuning-free fashion to conduct complex multimodal tasks. Instead of simple model composition, the i-Code Studio provides an integrative, flexible, and composable setting for developers to quickly and easily compose cutting-edge services and technologies tailored to their specific requirements. The i-Code Studio achieves impressive results on a variety of zero-shot multimodal tasks, such as video-to-text retrieval, speech-to-speech translation, and visual question answering. We also demonstrate how to quickly build a multimodal agent based on the i-Code Studio that can communicate and personalize for users.

The convergence of text, visual, and audio data is a key step towards human-like artificial intelligence, however the current Vision-Language-Speech landscape is dominated by encoder-only models which lack generative abilities. We propose closing this gap with i-Code V2, the first model capable of generating natural language from any combination of Vision, Language, and Speech data. i-Code V2 is an integrative system that leverages state-of-the-art single-modality encoders, combining their outputs with a new modality-fusing encoder in order to flexibly project combinations of modalities into a shared representational space. Next, language tokens are generated from these representations via an autoregressive decoder. The whole framework is pretrained end-to-end on a large collection of dual- and single-modality datasets using a novel text completion objective that can be generalized across arbitrary combinations of modalities. i-Code V2 matches or outperforms state-of-the-art single- and dual-modality baselines on 7 multimodal tasks, demonstrating the power of generative multimodal pretraining across a diversity of tasks and signals.

Learning tasks on source code (i.e., formal languages) have been considered recently, but most work has tried to transfer natural language methods and does not capitalize on the unique opportunities offered by code's known syntax. For example, long-range dependencies induced by using the same variable or function in distant locations are often not considered. We propose to use graphs to represent both the syntactic and semantic structure of code and use graph-based deep learning methods to learn to reason over program structures. In this work, we present how to construct graphs from source code and how to scale Gated Graph Neural Networks training to such large graphs. We evaluate our method on two tasks: VarNaming, in which a network attempts to predict the name of a variable given its usage, and VarMisuse, in which the network learns to reason about selecting the correct variable that should be used at a given program location. Our comparison to methods that use less structured program representations shows the advantages of modeling known structure, and suggests that our models learn to infer meaningful names and to solve the VarMisuse task in many cases. Additionally, our testing showed that VarMisuse identifies a number of bugs in mature open-source projects., Comment: Published in ICLR 2018. arXiv admin note: text overlap with arXiv:1705.07867

This paper presents a subject-independent facial action unit (AU) detection method by introducing the concept of relative AU detection, for scenarios where the neutral face is not provided. We propose a new classification objective function which analyzes the temporal neighborhood of the current frame to decide if the expression recently increased, decreased or showed no change. This approach is a significant change from the conventional absolute method which decides about AU classification using the current frame, without an explicit comparison with its neighboring frames. Our proposed method improves robustness to individual differences such as face scale and shape, age-related wrinkles, and transitions among expressions (e.g., lower intensity of expressions). Our experiments on three publicly available datasets (Extended Cohn-Kanade (CK+), Bosphorus, and DISFA databases) show significant improvement of our approach over conventional absolute techniques. Keywords: facial action coding system (FACS); relative facial action unit detection; temporal information, Comment: Accepted at IEEE Winter Conference on Applications of Computer Vision, Steamboat Springs Colorado, USA, 2014

Active Learning Method (ALM) is a soft computing method which is used for modeling and control, based on fuzzy logic. Although ALM has shown that it acts well in dynamic environments, its operators cannot support it very well in complex situations due to losing data. Thus ALM can find better membership functions if more appropriate operators be chosen for it. This paper substituted two new operators instead of ALM original ones; which consequently renewed finding membership functions in a way superior to conventional ALM. This new method is called Extended Active Learning Method (EALM)., Comment: 18 pages, 26 figures, 2 tables, submitted to the control engineering practice of Elsevier

The decision boundaries of Bayes classifier are optimal because they lead to maximum probability of correct decision. It means if we knew the prior probabilities and the class-conditional densities, we could design a classifier which gives the lowest probability of error. However, in classification based on nonparametric density estimation methods such as Parzen windows, the decision regions depend on the choice of parameters such as window width. Moreover, these methods suffer from curse of dimensionality of the feature space and small sample size problem which severely restricts their practical applications. In this paper, we address these problems by introducing a novel dimension reduction and classification method based on local component analysis. In this method, by adopting an iterative cross-validation algorithm, we simultaneously estimate the optimal transformation matrices (for dimension reduction) and classifier parameters based on local information. The proposed method can classify the data with complicated boundary and also alleviate the course of dimensionality dilemma. Experiments on real data show the superiority of the proposed algorithm in term of classification accuracies for pattern classification applications like age, facial expression and character recognition. Keywords: Bayes classifier, curse of dimensionality dilemma, Parzen window, pattern classification, subspace learning., Comment: This paper has been withdrawn by the author due to an error in experimental results

Active Learning Method (ALM) is a soft computing method used for modeling and control based on fuzzy logic. All operators defined for fuzzy sets must serve as either fuzzy S-norm or fuzzy T-norm. Despite being a powerful modeling method, ALM does not possess operators which serve as S-norms and T-norms which deprive it of a profound analytical expression/form. This paper introduces two new operators based on morphology which satisfy the following conditions: First, they serve as fuzzy S-norm and T-norm. Second, they satisfy Demorgans law, so they complement each other perfectly. These operators are investigated via three viewpoints: Mathematics, Geometry and fuzzy logic., Comment: 11 pages, 20 figures, under review of SPRINGER (Fuzzy Optimization and Decision Making)

In this paper a novel method called Extended Two-Dimensional PCA (E2DPCA) is proposed which is an extension to the original 2DPCA. We state that the covariance matrix of 2DPCA is equivalent to the average of the main diagonal of the covariance matrix of PCA. This implies that 2DPCA eliminates some covariance information that can be useful for recognition. E2DPCA instead of just using the main diagonal considers a radius of r diagonals around it and expands the averaging so as to include the covariance information within those diagonals. The parameter r unifies PCA and 2DPCA. r = 1 produces the covariance of 2DPCA, r = n that of PCA. Hence, by controlling r it is possible to control the trade-offs between recognition accuracy and energy compression (fewer coefficients), and between training and recognition complexity. Experiments on ORL face database show improvement in both recognition accuracy and recognition time over the original 2DPCA., Comment: Proc. of 4th International Conference on Intelligent Computer Communication and Processing (ICCP), Cluj-Napoca, Romania, pp. 295--298, 2008.

In this paper an accurate real-time sequence-based system for representation, recognition, interpretation, and analysis of the facial action units (AUs) and expressions is presented. Our system has the following characteristics: 1) employing adaptive-network-based fuzzy inference systems (ANFIS) and temporal information, we developed a classification scheme based on neuro-fuzzy modeling of the AU intensity, which is robust to intensity variations, 2) using both geometric and appearance-based features, and applying efficient dimension reduction techniques, our system is robust to illumination changes and it can represent the subtle changes as well as temporal information involved in formation of the facial expressions, and 3) by continuous values of intensity and employing top-down hierarchical rule-based classifiers, we can develop accurate human-interpretable AU-to-expression converters. Extensive experiments on Cohn-Kanade database show the superiority of the proposed method, in comparison with support vector machines, hidden Markov models, and neural network classifiers. Keywords: biased discriminant analysis (BDA), classifier design and evaluation, facial action units (AUs), hybrid learning, neuro-fuzzy modeling.

In this paper a novel efficient method for representation of facial action units by encoding an image sequence as a fourth-order tensor is presented. The multilinear tensor-based extension of the biased discriminant analysis (BDA) algorithm, called multilinear biased discriminant analysis (MBDA), is first proposed. Then, we apply the MBDA and two-dimensional BDA (2DBDA) algorithms, as the dimensionality reduction techniques, to Gabor representations and the geometric features of the input image sequence respectively. The proposed scheme can deal with the asymmetry between positive and negative samples as well as curse of dimensionality dilemma. Extensive experiments on Cohn-Kanade database show the superiority of the proposed method for representation of the subtle changes and the temporal information involved in formation of the facial expressions. As an accurate tool, this representation can be applied to many areas such as recognition of spontaneous and deliberate facial expressions, multi modal/media human computer interaction and lie detection efforts., Comment: Proc. of 16th Korea-Japan Joint Workshop on Frontiers of Computer Vision, Hiroshima, Japan, 2010.

Facial Action Coding System consists of 44 action units (AUs) and more than 7000 combinations. Hidden Markov models (HMMs) classifier has been used successfully to recognize facial action units (AUs) and expressions due to its ability to deal with AU dynamics. However, a separate HMM is necessary for each single AU and each AU combination. Since combinations of AU numbering in thousands, a more efficient method will be needed. In this paper an accurate real-time sequence-based system for representation and recognition of facial AUs is presented. Our system has the following characteristics: 1) employing a mixture of HMMs and neural network, we develop a novel accurate classifier, which can deal with AU dynamics, recognize subtle changes, and it is also robust to intensity variations, 2) although we use an HMM for each single AU only, by employing a neural network we can recognize each single and combination AU, and 3) using both geometric and appearance-based features, and applying efficient dimension reduction techniques, our system is robust to illumination changes and it can represent the temporal information involved in formation of the facial expressions. Extensive experiments on Cohn-Kanade database show the superiority of the proposed method, in comparison with other classifiers. Keywords: classifier design and evaluation, data fusion, facial action units (AUs), hidden Markov models (HMMs), neural network (NN).

In this paper, a novel method for representation and recognition of the facial expressions in two-dimensional image sequences is presented. We apply a variation of two-dimensional heteroscedastic linear discriminant analysis (2DHLDA) algorithm, as an efficient dimensionality reduction technique, to Gabor representation of the input sequence. 2DHLDA is an extension of the two-dimensional linear discriminant analysis (2DLDA) approach and it removes the equal within-class covariance. By applying 2DHLDA in two directions, we eliminate the correlations between both image columns and image rows. Then, we perform a one-dimensional LDA on the new features. This combined method can alleviate the small sample size problem and instability encountered by HLDA. Also, employing both geometric and appearance features and using an ensemble learning scheme based on data fusion, we create a classifier which can efficiently classify the facial expressions. The proposed method is robust to illumination changes and it can properly represent temporal information as well as subtle changes in facial muscles. We provide experiments on Cohn-Kanade database that show the superiority of the proposed method. KEYWORDS: two-dimensional heteroscedastic linear discriminant analysis (2DHLDA), subspace learning, facial expression analysis, Gabor wavelets, ensemble learning., Comment: This paper has been withdrawn by the author due to an error in experimental results