Search

Traditional phase I dose finding cancer clinical trial designs aim to determine the maximum tolerated dose (MTD) of the investigational cytotoxic agent based on a single toxicity outcome, assuming a monotone dose-response relationship. However, this assumption might not always hold for newly emerging therapies such as immuno-oncology therapies and molecularly targeted therapies, making conventional dose finding trial designs based on toxicity no longer appropriate. To tackle this issue, numerous early phase dose finding clinical trial designs have been developed to identify the optimal biological dose (OBD), which takes both toxicity and efficacy outcomes into account. In this article, we review the current model-assisted dose finding designs, BOIN-ET, BOIN12, UBI, TEPI-2, PRINTE, STEIN, and uTPI to identify the OBD and compare their operating characteristics. Extensive simulation studies and a case study using a CAR T-cell therapy phase I trial have been conducted to compare the performance of the aforementioned designs under different possible dose-response relationship scenarios. The simulation results demonstrate that the performance of different designs varies depending on the particular dose-response relationship and the specific metric considered. Based on our simulation results and practical considerations, STEIN, PRINTE, and BOIN12 outperform the other designs from different perspectives., Comment: This paper was recently accepted by the Journal of Biopharmaceutical Statistics in October 2024 and is currently undergoing proofing. The DOI for this paper is 10.1080/10543406.2024.2424845

When learning in strategic environments, a key question is whether agents can overcome uncertainty about their preferences to achieve outcomes they could have achieved absent any uncertainty. Can they do this solely through interactions with each other? We focus this question on the ability of agents to attain the value of their Stackelberg optimal strategy and study the impact of information asymmetry. We study repeated interactions in fully strategic environments where players' actions are decided based on learning algorithms that take into account their observed histories and knowledge of the game. We study the pure Nash equilibria (PNE) of a meta-game where players choose these algorithms as their actions. We demonstrate that if one player has perfect knowledge about the game, then any initial informational gap persists. That is, while there is always a PNE in which the informed agent achieves her Stackelberg value, there is a game where no PNE of the meta-game allows the partially informed player to achieve her Stackelberg value. On the other hand, if both players start with some uncertainty about the game, the quality of information alone does not determine which agent can achieve her Stackelberg value. In this case, the concept of information asymmetry becomes nuanced and depends on the game's structure. Overall, our findings suggest that repeated strategic interactions alone cannot facilitate learning effectively enough to earn an uninformed player her Stackelberg value.

We use $p$-rank bounds on partial ovoids and the classical bounds on Ramsey numbers to obtain upper bounds on the size of partial $m$-ovoids in finite classical polar spaces. These bounds imply non-existence of $m$-ovoids for new infinite families of polar spaces. We also give a probabilistic construction of large partial $m$-ovoids when $m$ grows linearly with the rank of the polar space. In the special case of the symplectic spaces over the binary field, we prove an equivalence between partial $m$-ovoids and a generalisation of Oddtown families from extremal set theory that has been studied under the name of $m$-nearly orthogonal sets. We give a new construction for large partial $2$-ovoids in these spaces and thus $2$-nearly orthogonal sets over the binary field. This construction uses triangle-free graphs associated to certain BCH codes whose complements have low $2$-rank and it gives an asymptotic improvement over the previous best construction. We give another construction of triangle-free graphs using a binary projective cap, which has low complementary rank over the reals. This improves the bounds in the recently introduced rank-Ramsey problem and it gives better constructions of large partial $m$-ovoids for $m > 2$ in the binary symplectic space., Comment: 13 pages, 1 Figure, added a new co-author and new results (Section 5.3)

We study price-discrimination games between buyers and a seller where privacy arises endogenously--that is, utility maximization yields equilibrium strategies where privacy occurs naturally. In this game, buyers with a high valuation for a good have an incentive to keep their valuation private, lest the seller charge them a higher price. This yields an equilibrium where some buyers will send a signal that misrepresents their type with some probability; we refer to this as buyer-induced privacy. When the seller is able to publicly commit to providing a certain privacy level, we find that their equilibrium response is to commit to ignore buyers' signals with some positive probability; we refer to this as seller-induced privacy. We then turn our attention to a repeated interaction setting where the game parameters are unknown and the seller cannot credibly commit to a level of seller-induced privacy. In this setting, players must learn strategies based on information revealed in past rounds. We find that, even without commitment ability, seller-induced privacy arises as a result of reputation building. We characterize the resulting seller-induced privacy and seller's utility under no-regret and no-policy-regret learning algorithms and verify these results through simulations., Comment: To appear in Symposium on Foundations of Responsible Computing (FORC 2024)

Motivated by the emergence of decentralized machine learning (ML) ecosystems, we study the delegation of data collection. Taking the field of contract theory as our starting point, we design optimal and near-optimal contracts that deal with two fundamental information asymmetries that arise in decentralized ML: uncertainty in the assessment of model quality and uncertainty regarding the optimal performance of any model. We show that a principal can cope with such asymmetry via simple linear contracts that achieve 1-1/e fraction of the optimal utility. To address the lack of a priori knowledge regarding the optimal performance, we give a convex program that can adaptively and efficiently compute the optimal contract. We also study linear contracts and derive the optimal utility in the more complex setting of multiple interactions.

The estimation of the direction of electromagnetic (EM) waves from a radio source using electrically short antennas is one of the challenging problems in the field of radio astronomy. In this paper we have developed an algorithm which performs better in direction and polarization estimations than the existing algorithms. Our proposed algorithm Snapshot Averaged Matrix Pencil Method (SAM) is a modification to the existing Matrix Pencil Method (MPM) based Direction of Arrival (DoA) algorithm. In general, MPM estimates DoA of the incoherent EM waves in the spectra using unitary transformations and least square method (LSM). Our proposed SAM modification is made in context to the proposed Space Electric and Magnetic Sensor (SEAMS) mission to study the radio universe below 16 MHz. SAM introduces a snapshot averaging method to improve the incoherent frequency estimation improving the accuracy of estimation. It can also detect polarization to differentiate between Right Hand Circular Polarlization (RHCP), Right Hand Elliptical Polarlization (RHEP), Left Hand Circular Polarlization (LHCP), Left Hand Elliptical Polarlization (LHEP) and Linear Polarlization (LP). This paper discusses the formalism of SAM and shows the initial results of a scaled version of a DoA experiment at a resonant frequency of ~72 MHz.

Magnetic bioactive glass ceramic (MGC) powders have been synthesized by sol gel route by systematically substituting silicon dioxide with iron oxide in the 45S5 glass composition. Powder x-ray diffraction studies revealed a variation in the percentage of combeite (Ca$_2$Na$_2$Si$_3$O$_9$), magnetite (Fe$_3$O$_4$), and hematite (Fe$_2$O$_3$) nanocrystalline phases in MGC powders as a function of composition. Zeta potential measurements showed that MGC containing up to 10 wt.% iron oxide formed stable suspensions. Saturation magnetization and heat generation capacity of MGC fluids increased with increase in iron oxide content. Degradation of MGC powders was investigated in phosphate buffer saline (PBS). In vitro bioactivity of the MGC powders taken in pellet form was confirmed by observing the pH variation as well as hydroxyapatite layer (HAp) formation upon soaking in modified simulated body fluid. These studies showed a decrement in overall bioactivity in samples with high iron oxide content due to the proportional decrease in silanol group. Monitoring the proliferation of MG-63 osteoblast cell in Dulbecco's Modified Eagle Medium (DMEM) revealed that MGC with up to 10 wt.% iron oxide exhibited acceptable viability. The systematic study revealed that the MGC with 10 wt.% iron oxide exhibited optimal cell viability, magnetic properties and induction heating capacity which were better than those of FluidMag-CT, which is used for hyperthermia treatment., Comment: 24 pages, 11 figures

Spoken Language Understanding (SLU) systems typically consist of a set of machine learning models that operate in conjunction to produce an SLU hypothesis. The generated hypothesis is then sent to downstream components for further action. However, it is desirable to discard an incorrect hypothesis before sending it downstream. In this work, we present two designs for SLU hypothesis rejection modules: (i) scheme R1 that performs rejection on domain specific SLU hypothesis and, (ii) scheme R2 that performs rejection on hypothesis generated from the overall SLU system. Hypothesis rejection modules in both schemes reject/accept a hypothesis based on features drawn from the utterance directed to the SLU system, the associated SLU hypothesis and SLU confidence score. Our experiments suggest that both the schemes yield similar results (scheme R1: 2.5% FRR @ 4.5% FAR, scheme R2: 2.5% FRR @ 4.6% FAR), with the best performing systems using all the available features. We argue that while either of the rejection schemes can be chosen over the other, they carry some inherent differences which need to be considered while making this choice. Additionally, we incorporate ASR features in the rejection module (obtaining an 1.9% FRR @ 3.8% FAR) and analyze the improvements., Comment: 5 pages. ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Research process automation -- the reliable, efficient, and reproducible execution of linked sets of actions on scientific instruments, computers, data stores, and other resources -- has emerged as an essential element of modern science. We report here on new services within the Globus research data management platform that enable the specification of diverse research processes as reusable sets of actions, \emph{flows}, and the execution of such flows in heterogeneous research environments. To support flows with broad spatial extent (e.g., from scientific instrument to remote data center) and temporal extent (from seconds to weeks), these Globus automation services feature: 1) cloud hosting for reliable execution of even long-lived flows despite sporadic failures; 2) a simple specification and extensible asynchronous action provider API, for defining and executing a wide variety of actions and flows involving heterogeneous resources; 3) an event-driven execution model for automating execution of flows in response to arbitrary events; and 4) a rich security model enabling authorization delegation mechanisms for secure execution of long-running actions across distributed resources. These services permit researchers to outsource and automate the management of a broad range of research tasks to a reliable, scalable, and secure cloud platform. We present use cases for Globus automation services, describe their design and implementation, present microbenchmark studies, and review experiences applying the services in a range of applications.

In this work, we demonstrate that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling (CLM) tasks, are more efficient few-shot learners than decoder-only models on various tasks. In particular, we train a 20 billion parameter multilingual seq2seq model called Alexa Teacher Model (AlexaTM 20B) and show that it achieves state-of-the-art (SOTA) performance on 1-shot summarization tasks, outperforming a much larger 540B PaLM decoder model. AlexaTM 20B also achieves SOTA in 1-shot machine translation, especially for low-resource languages, across almost all language pairs supported by the model (Arabic, English, French, German, Hindi, Italian, Japanese, Marathi, Portuguese, Spanish, Tamil, and Telugu) on Flores-101 dataset. We also show in zero-shot setting, AlexaTM 20B outperforms GPT3 (175B) on SuperGLUE and SQuADv2 datasets and provides SOTA performance on multilingual tasks such as XNLI, XCOPA, Paws-X, and XWinograd. Overall, our results present a compelling case for seq2seq models as a powerful alternative to decoder-only models for Large-scale Language Model (LLM) training.

We present results from a large-scale experiment on pretraining encoders with non-embedding parameter counts ranging from 700M to 9.3B, their subsequent distillation into smaller models ranging from 17M-170M parameters, and their application to the Natural Language Understanding (NLU) component of a virtual assistant system. Though we train using 70% spoken-form data, our teacher models perform comparably to XLM-R and mT5 when evaluated on the written-form Cross-lingual Natural Language Inference (XNLI) corpus. We perform a second stage of pretraining on our teacher models using in-domain data from our system, improving error rates by 3.86% relative for intent classification and 7.01% relative for slot filling. We find that even a 170M-parameter model distilled from our Stage 2 teacher model has 2.88% better intent classification and 7.69% better slot filling error rates when compared to the 2.3B-parameter teacher trained only on public data (Stage 1), emphasizing the importance of in-domain data for pretraining. When evaluated offline using labeled NLU data, our 17M-parameter Stage 2 distilled model outperforms both XLM-R Base (85M params) and DistillBERT (42M params) by 4.23% to 6.14%, respectively. Finally, we present results from a full virtual assistant experimentation platform, where we find that models trained using our pretraining and distillation pipeline outperform models distilled from 85M-parameter teachers by 3.74%-4.91% on an automatic measurement of full-system user dissatisfaction., Comment: KDD 2022