To overcome the high cost of data acquisition, few-shot semantic segmentation is studied to increase the training efficiency of limited data, but it fails to detect the narrow objects well. We find that the issue is caused by two main reasons: the enlarged receptive field of the baseline models and the high-proportional noisy labels of the narrow objects. An enlarged receptive field lets the model ignore detailed information that is important for the narrow objects, which can be affected by the same amount of noisy labels more critically than the large objects. To solve the issue, we propose a novel method to improve the performance of narrow objects in few-shot semantic segmentation. First of all, we diversify the size of the receptive field by extracting multiple prototypes from multi-level pyramidal feature maps, which is helpful to consider the detailed features of narrow objects. In addition, during model training, we simultaneously update uncertainty maps that determine the pixel-wise label reliability to detect and ignore noisy labels. We validate the proposed method, which shows impressive enhancement for narrow object segmentation both quantitatively and qualitatively over the prior research.

Sangwon Yu, Jongyoon Song, Bongkyu Hwang, Hoyoung Kang, Sooah Cho, Junhwa Choi, Seongho Joe, Taehee Lee, Youngjune Gwon, Sungroh Yoon. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Jongyoon Song, Nohil Park, Bongkyu Hwang, Jaewoong Yun, Seongho Joe, Youngjune Gwon, Sungroh Yoon. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

While message passing graph neural networks result in informative node embeddings, they may suffer from describing the topological properties of graphs. To this end, node filtration has been widely used as an attempt to obtain the topological information of a graph using persistence diagrams. However, these attempts have faced the problem of losing node embedding information, which in turn prevents them from providing a more expressive graph representation. To tackle this issue, we shift our focus to edge filtration and introduce a novel edge filtration-based persistence diagram, named Topological Edge Diagram (TED), which is mathematically proven to preserve node embedding information as well as contain additional topological information. To implement TED, we propose a neural network based algorithm, named Line Graph Vietoris-Rips (LGVR) Persistence Diagram, that extracts edge information by transforming a graph into its line graph. Through LGVR, we propose two model frameworks that can be applied to any message passing GNNs, and prove that they are strictly more powerful than Weisfeiler-Lehman type colorings. Finally we empirically validate superior performance of our models on several graph classification and regression benchmarks.

Contract documents can be modified just before signing, after the consensus, with the intention of defrauding the other party, which can have serious consequences for the deal. To prevent the issue, we propose a method to detect document changes between a scanned final document and its original electronic file using image-based comparison. Our method first finds the most appropriate augmentation for various document changes, such as rotations, contrast, ratio, or brightness changes which can occur while scanning documents. Then, we employ a hierarchical search strategy from large patches to small patches in a sliding window manner, which can reduce the computational complexity to compare all the details of the documents using the deep learning model. We built a new dataset of original-scanned document pair for the validation of our method. In the experiments, we show that our method outperforms the previous approaches using segmentation and character recognition models, even when the document suffers from both non-lingual and lingual changes.

The recent advances in representation learning inspire us to take on the challenging problem of unsupervised image classification tasks in a principled way. We propose ContraCluster, an unsupervised image classification method that combines clustering with the power of contrastive self-supervised learning. ContraCluster consists of three stages: (1) contrastive self-supervised pre-training (CPT), (2) contrastive prototype sampling (CPS), and (3) prototype-based semi-supervised fine-tuning (PB-SFT). CPS can select highly accurate, categorically prototypical images in an embedding space learned by contrastive learning. We use sampled prototypes as noisy labeled data to perform semi-supervised fine-tuning (PB-SFT), leveraging small prototypes and large unlabeled data to further enhance the accuracy. We demonstrate empirically that ContraCluster achieves new state-of-the-art results for standard benchmark datasets including CIFAR-10, STL-10, and ImageNet-10. For example, ContraCluster achieves about 90.8% accuracy for CIFAR-10, which outperforms DAC (52.2%), IIC (61.7%), and SCAN (87.6%) by a large margin. Without any labels, ContraCluster can achieve a 90.8% accuracy that is comparable to 95.8% by the best supervised counterpart.

We propose a self-supervised learning method for long text documents based on contrastive learning. A key to our method is Shuffle and Divide (SaD), a simple text augmentation algorithm that sets up a pretext task required for contrastive updates to BERT-based document embedding. SaD splits a document into two sub-documents containing randomly shuffled words in the entire documents. The sub-documents are considered positive examples, leaving all other documents in the corpus as negatives. After SaD, we repeat the contrastive update and clustering phases until convergence. It is naturally a time-consuming, cumbersome task to label text documents, and our method can help alleviate human efforts, which are most expensive resources in AI. We have empirically evaluated our method by performing unsupervised text classification on the 20 Newsgroups, Reuters-21578, BBC, and BBCSport datasets. In particular, our method pushes the current state-of-the-art, SS-SB-MT, on 20 Newsgroups by 20.94% in accuracy. We also achieve the state-of-the-art performance on Reuters-21578 and exceptionally-high accuracy performances (over 95%) for unsupervised classification on the BBC and BBCSport datasets.

Neural approaches for combinatorial optimization (CO) equip a learning mechanism to discover powerful heuristics for solving complex real-world problems. While neural approaches capable of high-quality solutions in a single shot are emerging, state-of-the-art approaches are often unable to take full advantage of the solving time available to them. In contrast, hand-crafted heuristics perform highly effective search well and exploit the computation time given to them, but contain heuristics that are difficult to adapt to a dataset being solved. With the goal of providing a powerful search procedure to neural CO approaches, we propose simulation-guided beam search (SGBS), which examines candidate solutions within a fixed-width tree search that both a neural net-learned policy and a simulation (rollout) identify as promising. We further hybridize SGBS with efficient active search (EAS), where SGBS enhances the quality of solutions backpropagated in EAS, and EAS improves the quality of the policy used in SGBS. We evaluate our methods on well-known CO benchmarks and show that SGBS significantly improves the quality of the solutions found under reasonable runtime assumptions.

The advancement in numerous generative models has a two-fold effect: a simple and easy generation of realistic synthesized images, but also an increased risk of malicious abuse of those images. Thus, it is important to develop a generalized detector for synthesized images of any GAN model or object category, including those unseen during the training phase. However, the conventional methods heavily depend on the training settings, which cause a dramatic decline in performance when tested with unknown domains. To resolve the issue and obtain a generalized detection ability, we propose Bilateral High-Pass Filters (BiHPF), which amplify the effect of the frequency-level artifacts that are generally found in the synthesized images of generative models. Also, to find the properties of the general frequency-level artifacts, we develop an additional method to adversarially extract the artifact compression map. Numerous experimental results validate that our method outperforms other state-of-the-art methods, even when tested with unseen domains.

Active Learning for discriminative models has largely been studied with the focus on individual samples, with less emphasis on how classes are distributed or which classes are hard to deal with. In this work, we show that this is harmful. We propose a method based on the Bayes’ rule, that can naturally incorporate class imbalance into the Active Learning framework. We derive that three terms should be considered together when estimating the probability of a classifier making a mistake for a given sample; i) probability of mislabelling a class, ii) likelihood of the data given a predicted class, and iii) the prior probability on the abundance of a predicted class. Implementing these terms requires a generative model and an intractable likelihood estimation. Therefore, we train a Variational Auto Encoder (VAE) for this purpose. To further tie the VAE with the classifier and facilitate VAE training, we use the classifiers’ deep feature representations as input to the VAE. By considering all three probabilities, among them, especially the data imbalance, we can substantially improve the potential of existing methods under limited data budget. We show that our method can be applied to classification tasks on multiple different datasets – including one that is a real-world dataset with heavy data imbalance – significantly outperforming the state of the art.

Singular value decomposition (SVD) is fundamentally important and broadly useful in both quantum and classical computing. There are several quantum algorithms known for SVD. Variational quantum algorithms with parametric quantum circuits (PQC) are the most promising approach to find SVD with near-term quantum computers. This paper reports a new method for quantum SVD (QSVD) that finds the singular vectors ordered by the magnitude of singular values by adding extra CNOT gates and a new local cost function design. This ordering is important because the singular vectors with the larger singular values is more representative for the given information. The ordering property of this approach is investigated with the standard Iris dataset by numerical simulations of 4-qubit states. Methods for choosing appropriate choices for the cost function hyperparameter and cost function terms as well as an application example of a quantum encoder are also discussed.

Denoising diffusion probabilistic models (DDPM) have shown remarkable performance in unconditional image generation. However, due to the stochasticity of the generative process in DDPM, it is challenging to generate images with the desired semantics. In this work, we propose Iterative Latent Variable Refinement (ILVR), a method to guide the generative process in DDPM to generate high-quality images based on a given reference image. Here, the refinement of the generative process in DDPM enables a single DDPM to sample images from various sets directed by the reference image. The proposed ILVR method generates high-quality images while controlling the generation. The controllability of our method allows adaptation of a single DDPM without any additional learning in various image generation tasks, such as generation from various downsampling factors, multi-domain image translation, paint-to-image, and editing with scribbles.

Contextualized word embeddings can lead to state-of-the-art performances in natural language understanding. Recently, a pre-trained deep contextualized text encoder such as BERT has shown its potential in improving natural language tasks including abstractive summarization. Existing approaches in dialogue summarization focus on incorporating a large language model into summarization task trained on large-scale corpora consisting of news articles rather than dialogues of multiple speakers. In this paper, we introduce self-supervised methods to compensate shortcomings to train a dialogue summarization model. Our principle is to detect incoherent information flows using pretext dialogue text to enhance BERT's ability to contextualize the dialogue text representations. We build and fine-tune an abstractive dialogue summarization model on a shared encoder-decoder architecture using the enhanced BERT. We empirically evaluate our abstractive dialogue summarizer with the SAMSum corpus, a recently introduced dataset with abstractive dialogue summaries. All of our methods have contributed improvements to abstractive summary measured in ROUGE scores. Through an extensive ablation study, we also present a sensitivity analysis to critical model hyperparameters, probabilities of switching utterances and masking interlocutors.

Machine Learning (ML) can help solve combinatorial optimization (CO) problems better. A popular approach is to use a neural net to compute on the parameters of a given CO problem and extract useful information that guides the search for good solutions. Many CO problems of practical importance can be specified in a matrix form of parameters quantifying the relationship between two groups of items. There is currently no neural net model, however, that takes in such matrix-style relationship data as an input. Consequently, these types of CO problems have been out of reach for ML engineers. In this paper, we introduce Matrix Encoding Network (MatNet) and show how conveniently it takes in and processes parameters of such complex CO problems. Using an end-to-end model based on MatNet, we solve asymmetric traveling salesman (ATSP) and flexible flow shop (FFSP) problems as the earliest neural approach. In particular, for a class of FFSP we have tested MatNet on, we demonstrate a far superior empirical performance to any methods (neural or not) known to date.

We describe an unsupervised domain adaptation framework for images by a transform to an abstract intermediate domain and ensemble classifiers seeking a consensus. The intermediate domain can be thought as a latent domain where both the source and target domains can be transferred easily. The proposed framework aligns both domains to the intermediate domain, which greatly improves the adaptation performance when the source and target domains are notably dissimilar. In addition, we propose an ensemble model trained by confusing multiple classifiers and letting them make a consensus alternately to enhance the adaptation performance for ambiguous samples. To estimate the hidden intermediate domain and the unknown labels of the target domain simultaneously, we develop a training algorithm using a double-structured architecture. We validate the proposed framework in hard adaptation scenarios with real-world datasets from simple synthetic domains to complex real-world domains. The proposed algorithm outperforms the previous state-of-the-art algorithms on various environments.

We propose DefogGAN, a generative approach to the problem of inferring state information hidden in the fog of war for real-time strategy (RTS) games. Given a partially observed state, DefogGAN generates defogged images of a game as predictive information. Such information can lead to create a strategic agent for the game. DefogGAN is a conditional GAN variant featuring pyramidal reconstruction loss to optimize on multiple feature resolution scales. We have validated DefogGAN empirically using a large dataset of professional StarCraft replays. Our results indicate that DefogGAN can predict the enemy buildings and combat units as accurately as professional players do and achieves a superior performance among state-of-the-art defoggers.

Contextualized representations from a pre-trained language model are central to achieve a high performance on downstream NLP task. The pre-trained BERT and A Lite BERT (ALBERT) models can be fine-tuned to give state-of-the-art results in sentence-pair regressions such as semantic textual similarity (STS) and natural language inference (NLI). Although BERT-based models yield the [CLS] token vector as a reasonable sentence embedding, the search for an optimal sentence embedding scheme remains an active research area in computational linguistics. This paper explores on sentence embedding models for BERT and ALBERT. In particular, we take a modified BERT network with siamese and triplet network structures called Sentence-BERT (SBERT) and replace BERT with ALBERT to create Sentence-ALBERT (SALBERT). We also experiment with an outer CNN sentence-embedding network for SBERT and SALBERT. We evaluate performances of all sentence-embedding models considered using the STS and NLI datasets. The empirical results indicate that our CNN architecture improves ALBERT models substantially more than BERT models for STS benchmark. Despite significantly fewer model parameters, ALBERT sentence embedding is highly competitive to BERT in downstream NLP evaluations.

In zero-shot cross-lingual transfer, a supervised NLP task trained on a corpus in one language is directly applicable to another language without any additional training. A source of cross-lingual transfer can be as straightforward as lexical overlap between languages (e.g., use of the same scripts, shared subwords) that naturally forces text embeddings to occupy a similar representation space. Recently introduced cross-lingual language model (XLM) pretraining brings out neural parameter sharing in Transformer-style networks as the most important factor for the transfer. In this paper, we aim to validate the hypothetically strong cross-lingual transfer properties induced by XLM pretraining. Particularly, we take XLM-RoBERTa (XLM-R) in our experiments that extend semantic textual similarity (STS), SQuAD and KorQuAD for machine reading comprehension, sentiment analysis, and alignment of sentence embeddings under various cross-lingual settings. Our results indicate that the presence of cross-lingual transfer is most pronounced in STS, sentiment analysis the next, and MRC the last. That is, the complexity of a downstream task softens the degree of cross-lingual transfer. All of our results are empirically observed and measured, and we make our code and data publicly available.

A Lite BERT (ALBERT) has been introduced to scale up deep bidirectional representation learning for natural languages. Due to the lack of pretrained ALBERT models for Korean language, the best available practice is the multilingual model or resorting back to the any other BERT-based model. In this paper, we develop and pretrain KoreALBERT, a monolingual ALBERT model specifically for Korean language understanding. We introduce a new training objective, namely Word Order Prediction (WOP), and use alongside the existing MLM and SOP criteria to the same architecture and model parameters. Despite having significantly fewer model parameters (thus, quicker to train), our pretrained KoreALBERT outperforms its BERT counterpart on 6 different NLU tasks. Consistent with the empirical results in English by Lan et al., KoreALBERT seems to improve downstream task performance involving multi-sentence encoding for Korean language. The pretrained KoreALBERT is publicly available to encourage research and application development for Korean NLP.

In neural combinatorial optimization (CO), reinforcement learning (RL) can turn a deep neural net into a fast, powerful heuristic solver of NP-hard problems. This approach has a great potential in practical applications because it allows near-optimal solutions to be found without expert guides armed with substantial domain knowledge. We introduce Policy Optimization with Multiple Optima (POMO), an end-to-end approach for building such a heuristic solver. POMO is applicable to a wide range of CO problems. It is designed to exploit the symmetries in the representation of a CO solution. POMO uses a modified REINFORCE algorithm that forces diverse rollouts towards all optimal solutions. Empirically, the low-variance baseline of POMO makes RL training fast and stable, and it is more resistant to local minima compared to previous approaches. We also introduce a new augmentation-based inference method, which accompanies POMO nicely. We demonstrate the effectiveness of POMO by solving three popular NP-hard problems, namely, traveling salesman (TSP), capacitated vehicle routing (CVRP), and 0-1 knapsack (KP). For all three, our solver based on POMO shows a significant improvement in performance over all recent learned heuristics. In particular, we achieve the optimality gap of 0.14% with TSP100 while reducing inference time by more than an order of magnitude.

We describe a new approach that improves the training of generative adversarial nets (GANs) for synthesizing diverse images from a text input. Our approach is based on the conditional version of GANs and expands on previous work leveraging an auxiliary task in the discriminator. Our generated images are not limited to certain classes and do not suffer from mode collapse while semantically matching the text input. A key to our training methods is how to form positive and negative training examples with respect to the class label of a given image. Instead of selecting random training examples, we perform negative sampling based on the semantic distance from a positive example in the class. We evaluate our approach using the Oxford-102 flower dataset, adopting the inception score and multi-scale structural similarity index (MS-SSIM) metrics to assess discriminability and diversity of the generated images. The empirical results indicate greater diversity in the generated images, especially when we gradually select more negative training examples closer to a positive example in the semantic space.

We present a clustering-based language model using word embeddings for text readability prediction. Presumably, an Euclidean semantic space hypothesis holds true for word embeddings whose training is done by observing word co-occurrences. We argue that clustering with word embeddings in the metric space should yield feature representations in a higher semantic space appropriate for text regression. Also, by representing features in terms of histograms, our approach can naturally address documents of varying lengths. An empirical evaluation using the Common Core Standards corpus reveals that the features formed on our clustering-based language model significantly improve the previously known results for the same corpus in readability prediction. We also evaluate the task of sentence matching based on semantic relatedness using the Wiki-SimpleWiki corpus and find that our features lead to superior matching performance.

We propose a novel RF signal classification method based on sparse coding, an unsupervised learning method popular in computer vision. In particular, we employ a convolutional sparse coder that can extract high-level features of an unknown received signal by maximal similarity matching against an over-complete dictionary of filter patterns. Such dictionary can be either generated or learned in an unsupervised fashion from measured signal examples conveying no ground-truth labels. The computed sparse code is then applied to train SVM classifiers for discriminating RF signals. As a result, the proposed approach can achieve blind signal classification that requires no prior knowledge (e.g., MCS, pulse shaping) about the signals present in an arbitrary RF channel. Since modulated RF signals undergo pulse shaping to aid the matched filter detection, our method exploits variability in relative similarity against the dictionary atoms as the key discriminating factor for classification. Our experimental results indicate that we can blindly separate different classes of digitally modulated signals with a 0.703 recall and 0.246 false alarm at 20dB SNR. Provided a small labeled dataset for supervised classifier training, we could improve the classification performance to a 0.878 recall and 0.141 false alarm.

We present Deep Sparse-coded Network (DSN), a deep architecture based on multilayer sparse coding. It has been considered difficult to learn a useful feature hierarchy by stacking sparse coding layers in a straightforward manner. The primary reason is the modeling assumption for sparse coding that takes in a dense input and yields a sparse output vector. Applying a sparse coding layer on the output of another tends to violate the modeling assumption. We overcome this shortcoming by interlacing nonlinear pooling units. Average- or max-pooled sparse codes are aggregated to form dense input vectors for the next sparse coding layer. Pooling achieves nonlinear activation analogous to neural networks while not introducing diminished gradient flows during the training. We introduce a novel backpropagation algorithm to finetune the proposed DSN beyond the pretraining via greedy layerwise sparse coding and dictionary learning. We build an experimental 4-layer DSN with the ℓ1-regularized LARS and the greedy-ℓ0OMP, and demonstrate superior performance over a similarly-configured stacked autoencoder (SAE) on CIFAR-10.

Spoken language recognition requires a series of signal processing steps and learning algorithms to model distinguishing characteristics of different languages. In this paper, we present a sparse discriminative feature learning framework for language recognition. We use sparse coding, an unsupervised method, to compute efficient representations for spectral features from a speech utterance while learning basis vectors for language models. Differentiated from existing approaches in sparse representation classification, we introduce a maximum a posteriori (MAP) adaptation scheme based on online learning that further optimizes the discriminative quality of sparse-coded speech features. We empirically validate the effectiveness of our approach using the NIST LRE 2015 dataset.

Competing Cognitive Radio Network (CCRN) coalesces communicator (comm) nodes and jammers to achieve maximal networking efficiency against adversarial threats. We have previously developed two contrasting approaches based on multiarmed bandit (MAB) and value-iterated Q-learning. Despite their differences, both approaches have demonstrated the efficacy of applying a machine learning technique to jointly compute comm and jammer actions in hypothetical two-network competition for an open dynamic spectrum. When sampled channel reward characteristics are time-invariant-i.e., stationarity of learned information, both MAB and Q-learning based strategies have resulted in the best possible reward empirically.

We propose a data-driven geolocation method on microblog text. Key idea underlying our approach is sparse coding, an unsupervised learning algorithm. Unlike conventional positioning algorithms, we geolocate a user by identifying features extracted from her social media text. We also present an enhancement robust to a random erasure of words in the text and report our experimental results with uniformly or randomly subsampled microblog text. Our solution features a novel two-step procedure consisting of upconversion and iterative refinement by joint sparse coding. As a result, we can reduce the computational cost of geolocation while preserving accuracy. In the light of information preservation and privacy, we remark potential applications of this paper.

This paper describes an adaptation of cognitive radio technology for tactical wireless networking. We introduce Competing Cognitive Radio Network (CCRN) featuring both communicator and jamming cognitive radio nodes that strategize in taking actions on an open spectrum under the presence of adversarial threats. We present the problem in the Multi-armed Bandit (MAB) framework and develop the optimal media access strategy consisting of mixed communicator and jammer actions in a Bayesian setting for Thompson sampling based on extreme value theory. Empirical results are promising that the proposed strategy seems to outperform Lai & Robbins and UCB, some of the most important MAB algorithms known to date.

We propose a spectrum analyzer that leverages many networked commodity sensor nodes, each of which samples its portion in a wideband spectrum. The sensors operate in parallel and transmit their measurements over a wireless network without performing any significant computations such as FFT. The measurements are forwarded to the backend of the system where spectrum analysis takes place. In particular, we propose a solution that compresses the raw measurements in a simple random linear projection and combines the compressed measurements from multiple sensors in-network. As a result, we achieve a substantial reduction in the network bandwidth requirement to operate the proposed system. We discover that the overall communication cost can be independent of the number of sensors and is affected only by sparsity of discretized spectrum under analysis. This principle founds the basis for a claim that our network-based spectrum analyzer can scale up the number of sensor nodes to process a very wide spectrum block potentially having a GHz bandwidth. We devise a novel recovery algorithm that systematically undoes compressive encoding and in-network combining done to the raw measurements, incorporating the least squares and I1-minimization decoding used in compressive sensing, and demonstrate that the algorithm can effectively restore an accurate estimate of the original data suitable for finegrained spectrum analysis. We present mathematical analysis and empirical evaluation of the system with software-defined radios.

We describe a method of integrating Karhunen-Loève Transform (KLT) into compressive sensing, which can as a result improve the compression ratio without affecting the accuracy of decoding. We present two complementary results: 1) by using KLT to find an optimal basis for decoding we can drastically reduce the number of measurements for compressive sensing used in applications such as radio spectrum analysis; 2) by using compressive sensing we can estimate and recover the KLT basis from compressive measurements of an input signal. In particular, we propose CS-KLT, an online estimation algorithm to cope with nonstationarity of wireless channels in reality. We validate our results with empirical data collected from a wideband UHF spectrum and field experiments to detect multiple radio transmitters, using software-defined radios.

We present statistical screening of test vectors for detecting a Trojan, malicious circuitry hidden inside an integrated circuit (IC). When applied a test vector, a Trojan-embedded chip draws extra leakage current that is unfortunately too small for the detector in most cases and concealed by process variation related to chip fabrication. To remedy the problem, we formulate a statistical approach that can screen and select test vectors in detecting Trojans. We validate our approach analytically and with gate-level simulations and show that our screening method leads to a substantial reduction in false positives and false negatives when detecting IC Trojans of various sizes.

We propose a gaining architecture, called the overlay server system (OSS), for supporting multiparty online games over mobile networks. In OSS architecture, overlay server nodes (OSN) are distributed across the core network architecture of the mobile operator and each node is responsible for both running the game applications and performing the overlay routing. The architecture allows third party game servers that are located outside the core network of the mobile operator to push the execution of delay and bandwidth constrained game components into the core to achieve better quality of service (QoS). When a game consists of multiple mutually exclusive game objects such that each object can be maintained independently, OSS allows per object QoS optimization via selecting the jointly optimal location (i.e. OSN) and overlay routes for each object. This fine-grain optimization benefits most when different objects have different QoS requirements and they are accessed by different users. OSS dynamically adapts to the changes in game and network conditions by switching to better OSN and overlay routes. Our performance analysis on different core network topologies and usage patterns demonstrates that OSS has significant advantages over the alternative peer to peer (P2P), proxy-server, and client-server architectures. The analysis also underlines the individual contributions of object placement and jointly performed overlay route optimization to the performance gain.

Fast mobile IPv6 (FMIP) and hierarchical mobile IPv6 (HMIP) are enhancements to the standard mobile IPv6 (SMIP) protocol for reducing handover latency and data loss, and for localized mobility management. In this paper, we present scalability and robustness analysis of the three protocols and two hybrid protocols of FMIP and HMIP, using a large-scale simulation. The simulation results indicate that FMIP achieves the best handover performance. HMIP incurs considerably less per-handover signaling overhead than FMIP on the wireless link, but HMIP data traffic has a fixed and permanent overhead even after handover. Hybrid protocols achieve FMIP-like handover performance and improve handover signaling overhead but cannot remove tunneling overhead. Hybrid protocols are also more robust to access router and home agent failures.

We present algorithms for estimating the location of stationary and mobile users based on heterogeneous indoor RF technologies. We propose two location algorithms, selective fusion location estimation (SELFLOC) and region of confidence (RoC), which can be used in conjunction with classical location algorithms such as triangulation, or with third-party commercial location estimation systems. The SELFLOC algorithm infers the user location by selectively fusing location information from multiple wireless technologies and/or multiple classical location algorithms in a theoretically optimal manner. The RoC algorithm attempts to overcome the problem of aliasing in the signal domain, where different physical locations have similar RF characteristics, which is particularly acute when users are mobile. We have empirically validated the proposed algorithms using wireless LAN and Bluetooth technology. Our experimental results show that applying SELFLOC for stationary users when using multiple wireless technologies and multiple classical location algorithms can improve location accuracy significantly, with mean distance errors as low as 1.6 m. For mobile users we find that using RoC can allow us to obtain mean errors as low as 3.7 m. Both algorithms can be used in conjunction with a commercial location estimation system and improve its accuracy further.

We introduce enhanced forwarding from the previous care-of address (EFWD) for fast handovers in mobile IPv6. EFWD enables a mobile node to directly control a bi-directional tunnel that is used to redirect data from the previous subnet's access router to new subnet's access router. As a result, data loss during a handover can be prevented. EFWD also reduces handover latency by expediting the mobile node's movement detection and new subnet's access router discovery, and by eliminating time to acquire a new care-of address. Main benefit of EFWD is the removal of link layer pre-triggers that are required by fast mobile IPv6 and generally infeasible for wireless technologies other than cellular systems. Empirical results indicate that performance of EFWD is nearly optimal as fast mobile IPv6, which incurs 30-40 msec handover latency with 3-4 lost UDP packets.

We investigate fast IP handoffs in wireless LAN networks. As a simple mobile-controlled approach, we introduce mobile initiated tunneling handoff protocol for IPv4 (MITHv4). Our experimental results show that MITHv4 can achieve optimized low latency and low loss IP handoffs in wireless LAN networks. Furthermore, MITHv4 significantly reduces the link layer trigger requirements and substantial access network support to synchronize link layer and IP layer handoffs that fast Mobile IPv4 (FMIPv4) protocols heavily rely on. These benefits of MITHv4 are crucial for wireless LAN networks where the required link layer triggers for FMIPv4 are not feasible due to limited access network control.