nips nips2012 nips2012-46 knowledge-graph by maker-knowledge-mining
Source: pdf
Author: Ognjen Arandjelovic
Abstract: The interaction between the patient’s expected outcome of an intervention and the inherent effects of that intervention can have extraordinary effects. Thus in clinical trials an effort is made to conceal the nature of the administered intervention from the participants in the trial i.e. to blind it. Yet, in practice perfect blinding is impossible to ensure or even verify. The current standard is follow up the trial with an auxiliary questionnaire, which allows trial participants to express their belief concerning the assigned intervention and which is used to compute a measure of the extent of blinding in the trial. If the estimated extent of blinding exceeds a threshold the trial is deemed sufficiently blinded; otherwise, the trial is deemed to have failed. In this paper we make several important contributions. Firstly, we identify a series of fundamental problems of the aforesaid practice and discuss them in context of the most commonly used blinding measures. Secondly, motivated by the highlighted problems, we formulate a novel method for handling imperfectly blinded trials. We too adopt a post-trial feedback questionnaire but interpret the collected data using an original approach, fundamentally different from those previously proposed. Unlike previous approaches, ours is void of any ad hoc free parameters, is robust to small changes in auxiliary data and is not predicated on any strong assumptions used to interpret participants’ feedback. 1
Reference: text
sentIndex sentText sentNum sentScore
1 Assessing Blinding in Clinical Trials Ognjen Arandjelovi´ c Deakin University, Australia Abstract The interaction between the patient’s expected outcome of an intervention and the inherent effects of that intervention can have extraordinary effects. [sent-1, score-0.43]
2 Thus in clinical trials an effort is made to conceal the nature of the administered intervention from the participants in the trial i. [sent-2, score-1.01]
3 Yet, in practice perfect blinding is impossible to ensure or even verify. [sent-5, score-0.627]
4 The current standard is follow up the trial with an auxiliary questionnaire, which allows trial participants to express their belief concerning the assigned intervention and which is used to compute a measure of the extent of blinding in the trial. [sent-6, score-1.9]
5 If the estimated extent of blinding exceeds a threshold the trial is deemed sufficiently blinded; otherwise, the trial is deemed to have failed. [sent-7, score-1.186]
6 Firstly, we identify a series of fundamental problems of the aforesaid practice and discuss them in context of the most commonly used blinding measures. [sent-9, score-0.645]
7 1 Introduction Ultimately, the main aim of a clinical trial is straightforward: it is to examine and quantify the effectiveness of a treatment of interest. [sent-13, score-0.706]
8 Blinding refers to the concealment of the type of administered intervention from the individuals/patients participating in a trial and its aim is to eliminate differential placebo effect between groups [10, 3, 11]. [sent-18, score-0.659]
9 Although conceptually simple, the problem of blinding poses difficult challenges in a practical clinical setup. [sent-19, score-0.669]
10 The first of these stems from the difficulty of ensuring that absolute blinding with respect to a particular trial variable is achieved. [sent-21, score-0.858]
11 The second challenge arises as a consequence of the fact that blinding can only be attempted with respect to those variables of the trial which have been identified as revealing of the treatment administered. [sent-22, score-1.14]
12 Put differently, it is always possible that a particular variable which can reveal the nature of the treatment to a trial participant is not identified by the trial designers and thus that no blinding with respect to it is attempted or achieved. [sent-23, score-1.506]
13 Given that it is both practically and in principle impossible to ensure perfect blinding, the practice of post hoc assessment of the level of blinding achieved has been gaining popularity and general acceptance by the clinical community. [sent-25, score-0.787]
14 Motivated by these, in the present work we propose a novel statistical framework and use it to derive an original method for integrated trial assessment which is experimentally shown to produce more meaningful and more clearly interpretable data. [sent-29, score-0.342]
15 1 Method 1: James’s Blinding Index At the heart of the so-called blinding index proposed by James et al. [sent-33, score-0.667]
16 [7] is the observation that the effect of a particular intervention is affected by the participant’s perception of the effectiveness of the intervention the participant believes was administered. [sent-34, score-0.504]
17 For example, a control group member who incorrectly believes to be a member of the treatment group may indeed experience positive effects expected from the studied treatment. [sent-35, score-0.556]
18 propose the use of a post-trial questionnaire to assess the level of blinding in a trial. [sent-38, score-0.673]
19 The participants are asked if they believe that they were assigned to the (i) control or (ii) treatment groups, or (iii) if they are uncertain of their assignment (the “don’t know” response). [sent-39, score-0.854]
20 allowing the participants to quantify the strength of their belief. [sent-42, score-0.418]
21 Blinding Level The existing work on the assessment of trial blinding uses the collected auxiliary data to calculate a statistic referred to as the blinding index. [sent-43, score-1.607]
22 Thus ρ1 = 1 indicates perfect blinding and ρ1 = 0 an unblinded trial. [sent-46, score-0.73]
23 It is readily apparent that ρ1 is a concave function which attains its maximal value of 1 when (i) all participants are uncertain of their assignment or (ii) when all participants have an incorrect belief regarding their assignment. [sent-49, score-1.03]
24 While it is tempting to reason that blinding must have been successful since no participant correctly guessed their assignment, it would be erroneous to do so. [sent-51, score-0.668]
25 In particular, the consistency of the wrong belief amongst trial participants actually reveals unblinding, but with the participants’ incorrect association of the unblinded factor with the corresponding group assignment. [sent-52, score-0.959]
26 For example, the treatment may cause perceivable side effects (thus unblinding the participants) and the worsening of the condition of the treatment group participants. [sent-53, score-0.768]
27 2 Method 2: Bang’s Blinding Index The blinding index ρ1 places a lot of value on those participants who plead ignorance regarding their assignment status. [sent-56, score-1.146]
28 argue that the non-decisive “don’t know” response may not express a 2 (a) (b) Figure 1: Dependency of the blinding indexes (a) ρ1 and (b) ρ2 on the proportions of “don’t know” responses P0 , and the correct assignment guesses PT + and PC− . [sent-58, score-0.857]
29 Because decisive responses can be in either the positive or the negative direction, the index is asymmetrical and can be applied separately to treatment and control groups. [sent-62, score-0.498]
30 For a 3-tier auxiliary questionnaire, the index for the treatment group is defined as: ρ2 = 2 PC− P T − + PT + −1 · . [sent-63, score-0.495]
31 5 an unblinded trial with incorrect assignment association, as discussed in Sec 2. [sent-71, score-0.472]
32 As the plot shows, this index achieves its perfect blinding value only when P0 = 1. [sent-73, score-0.673]
33 Also, PT + = PC− = 1 and P0 = 0 deems the trial unblinded, as does PT + = PC− = 0 and P0 = 0 but with the incorrect assignment association. [sent-75, score-0.408]
34 3 Limitations of the Current Best Standards In the preceding sections we described two blinding indexes most widely used in practice to assess the level of blinding in controlled clinical trials. [sent-77, score-1.346]
35 Adjustment of Free Parameters One of the most obvious difficulties encountered when applying either of the described blinding indexes concerns the need to choose appropriate values for the free parameters in Equations (2) and (3) in their general form. [sent-79, score-0.677]
36 use the same type of feedback data collected from the trial participants – the participants’ stated belief regarding their trial group assignment and their degree of confidence. [sent-88, score-1.28]
37 interpret the non-decisive, “don’t know” response as indicative of true lack of knowledge regarding the nature of the intervention (treatment or control). [sent-91, score-0.321]
38 If the trial participants are ignorant of their group assignment, it is assumed that they have indeed been blinded. [sent-92, score-0.753]
39 Instead, this response may be seen as a conservative one, reflecting the participants’ desire to appear balanced in their judgement or indeed the response that the participants believe would please the trial administration staff. [sent-95, score-0.806]
40 Thus, ρ2 mostly relies on the responses of those trial participants who did express belief regarding their group assignment. [sent-96, score-0.905]
41 As Hemili¨ amongst others notes, because the participants’ feedback is collected post a hoc it is possible that even a perfectly blinded subject becomes aware of the correct assignment by virtue of observing the effects (or lack thereof) of the assigned intervention [5]. [sent-99, score-0.583]
42 However, this is in most cases unsatisfactory as the participants would not have yet been exposed to any unblinded aspects of the trial. [sent-101, score-0.501]
43 establish the level of blindness in a trial by computing a blinding index and then comparing it with a predefined threshold. [sent-105, score-0.956]
44 This hard thresholding whereby a trial is considered either sufficiently well blinded or not means that the outcome of the blinding assessment can exhibit high sensitivity to small differences in participants’ responses. [sent-106, score-1.103]
45 Yet, such binarization in some form is necessitated by the nature of the blinding indexes because neither of the two described statistics has a clear practical interpretation in the clinical context. [sent-108, score-0.811]
46 Specifically, observe that the analysis of the trial outcome data is separated from the blinding assessment. [sent-111, score-0.922]
47 Indeed, only if the trial is deemed sufficiently well blinded does the analysis of actual trial data proceed. [sent-112, score-0.681]
48 Thus, if the blinding index falls short of the predetermined threshold, the data is effectively thrown away and the trial needs to be repeated. [sent-113, score-0.904]
49 On the other hand, if the blinding index exceeds the threshold, the analysis of data is performed in the same manner regardless of the actual value of the index, that is, regardless of whether it is just above the threshold or if it indicates perfect blinding. [sent-114, score-0.673]
50 The variety of problems that emerges from the atomization of different statistical aspects of a trial is inherently rooted in the very nature of the framework adopted by James et al. [sent-115, score-0.365]
51 For example, neither tells the clinician the probability that a particular portion of the participants were unblinded, nor the probability of a particular level of unblinding. [sent-119, score-0.455]
52 Instead, from the point of view of a clinician, the blinding index behaves like a black box which deems the trial well blinded or not, with little additional insight. [sent-120, score-1.072]
53 In the general case, the effectiveness of a particular intervention in a trial participant depends on the inherent effects of the intervention, as well as the participant’s expectations (conscious or not). [sent-127, score-0.61]
54 Thus, in the interpretation of trial results, we separately consider each population of participants which share the same combination of the type of intervention and the expressed belief regarding this group assignment. [sent-128, score-1.052]
55 A key idea of the proposed method is that because the outcome of an intervention depends on both the inherent effects of the intervention and the participants’ expectations, the effectiveness should be inferred in a like-for-like fashion. [sent-130, score-0.493]
56 In other words, the response observed in, say, the sub-group of participants assigned to the control group whose feedback professes belief in the control group 4 Figure 2: Conceptual illusProbability density tration of the proposed statistical model for the 3-tier feedback questionnaire. [sent-131, score-0.963]
57 Dotted and solid lines show respectively the probability density functions of the measured trial outcome across individuals in the three control and treatment sub-groups. [sent-132, score-0.745]
58 Outcome magnitude assignment should be compared with the response of only the sub-group of the treatment group who equally professed belief in the control group assignment. [sent-133, score-0.695]
59 Similarly, the “don’t know” sub-groups should be compared only with each other, as should the subgroups corresponding to the belief in the treatment assignment. [sent-134, score-0.441]
60 , xCg } be the trial outcome data collected from a control sub-group and (n (1) ) DT g = {xT g , . [sent-145, score-0.405]
61 Then, if Dg = DCg ∪ DT g is the totality of all data of participants who believe they were assigned to the group g: p(∆x | Dg ) = P (Dg | ∆x) p(∆x) . [sent-149, score-0.544]
62 However, as we will discuss in Sec 6, both Eq (7) and (9) have significance in the interpretation of trial results and their joint consideration can be used to reveal important additional information about the effectiveness of the treatment. [sent-154, score-0.36]
63 Exp 1: Reference For our first experiment, we simulated a trial involving 200 individuals, half of which were assigned to the control and half to the treatment group. [sent-167, score-0.652]
64 For each of the groups, 60% of the participants were taken to be in the “undecided” subgroups GC0 and GT 0 . [sent-168, score-0.489]
65 The remaining 40% of the participants was split between correct and incorrect guesses of the assigned intervention in proportion 3 : 1. [sent-169, score-0.698]
66 In this initial experiment we assume that all participants correctly disclosed their belief regarding which group they were assigned to. [sent-170, score-0.678]
67 We set the differential effect of treatment to ∆x = 0. [sent-172, score-0.355]
68 Relative to genuine lack of belief in either control or treatment group assignments, belief in control group assignment was set to exhibit negative effect of magnitude 0. [sent-175, score-0.821]
69 2 and that in treatment group assignment a positive effect of magnitude 0. [sent-176, score-0.468]
70 The maximum a posteriori (MAP) value of the estimate of the differential effectiveness of the treatment is ∆x∗ ≈ 0. [sent-181, score-0.391]
71 In comparison, when the differential effectiveness is estimated by subtracting the mean response of the control group from that of the treatment group, without the use of our matching sub-groups based statistical model, the estimate is ∆x ≈ 0. [sent-184, score-0.577]
72 Finally, the corresponding values of the blinding indices proposed by James et al. [sent-186, score-0.621]
73 Notice that the former indicates a level of blinding roughly half way between a perfectly blinded and unblinded trial, while the latter deems the trial nearly perfectly blinded. [sent-191, score-1.129]
74 1 Exp 2: Conservative Distortion We modify the baseline experiment by simulating conservative behavioural tendency of participants in a trial. [sent-194, score-0.421]
75 This was achieved by randomly choosing individuals from decisive subgroups and re-assigning them to their corresponding indecisive subgroup without changing their treatment’s observed effectiveness. [sent-195, score-0.313]
76 Fig 3(b) also shows the semi-amalgamated posterior obtained using only decisive subgroups which, by experimental design, 6 comprise data of only those individuals which honestly disclosed their belief of group assignment. [sent-201, score-0.442]
77 In Sec 6 we will show how the difference in statistical features of sub-group posteriors can be used to select the most reliable posteriors to amalgamate, as well as to reveal additional insight into the nature of the studied treatment and the blinding in the trial. [sent-205, score-1.098]
78 (a) Sub-group posteriors (b) Full posteriors Figure 3: Exp 2: (a) Posteriors for the differential effect treatment computed using the data Dg of each experimental sub-group comprising control and treatment individuals matched by their feedback. [sent-206, score-0.919]
79 (b) Posterior for the differential effect treatment computed using all available data. [sent-207, score-0.355]
80 Exp 3: Asymmetric Progressive Unblinding Starting with the baseline setup, we simulate unblinding of previously undecided individuals of the treatment group. [sent-208, score-0.476]
81 The robustness of our method is illustrated in Fig 4(a), which shows the MAP estimate of the effectiveness of the treatment after an increasing number of participants were unblinded. [sent-210, score-0.743]
82 The plots in Fig 4(b) show the variation of the two blinding indexes throughout the experiment. [sent-213, score-0.677]
83 (a) (b) Figure 4: Exp 3: (a) The MAP estimate of the treatment effectiveness as the participants assigned to the treatment group are progressively unblinded. [sent-218, score-1.171]
84 (b) The values of the blinding indexes ρ1 (blue line) and ρ2 (red line), computed at each step of the progressive unblinding of the participants assigned to the treatment group. [sent-219, score-1.541]
85 Exp 4: Symmetric Progressive Unblinding As in Exp 3 we start with the baseline setup and simulate unblinding of previously undecided individuals of the treatment group. [sent-220, score-0.476]
86 We illustrate the robustness of the method by plotting the MAP estimate of the effectiveness of the treatment in Fig 5(a). [sent-222, score-0.345]
87 As before, the estimate only shows small random perturbations, as expected in any experiment with a stochastic nature and is to be contrasted with the plots in Fig 5(b) which show the changes in the two blinding indexes throughout the experiment. [sent-223, score-0.703]
88 7 (a) (b) Figure 5: Exp 4: (a) The MAP estimate of the treatment effectiveness as the participants assigned to both the treatment and the control groups are progressively unblinded. [sent-228, score-1.151]
89 (b) The values of the blinding indexes ρ1 (blue line) and ρ2 (red line), computed at each step of the progressive unblinding. [sent-229, score-0.715]
90 In other words, it is possible that none of the participants of the treatment or the control group expressed a particular belief regarding their treatment assignment. [sent-233, score-1.209]
91 Firstly, note that whenever at least one pair of matching sub-groups is non-empty, the proposed method is able to compute a meaningful estimate of differential treatment effectiveness. [sent-235, score-0.353]
92 The absence of individuals in GT + may indicate that the participants assigned to the treatment group have either been poorly blinded but misidentified the received treatment, or that the treatment was vastly ineffective and was recognized as such by the participants assigned to it. [sent-237, score-1.756]
93 Similarly, the absence of individuals in GT − may indicate that the participants assigned to the treatment group have either been poorly blinded and correctly identified the received treatment, or that the treatment was obviously effective. [sent-238, score-1.302]
94 The cause of degeneration can then be determined using the knowledge of the administered interventions, and the statistics of both auxiliary responses and trial outcomes. [sent-240, score-0.45]
95 While the latter of these is of primary interest, the clinician can derive further useful insight into the nature of studied treatment by comparative examination of sub-group posteriors too. [sent-243, score-0.478]
96 individuals which did not respond positively to the treatment which in most people does produce a positive result [4, 8]. [sent-251, score-0.347]
97 We demonstrated a series of fundamental flaws in blinding index based approaches and thus proposed a novel framework. [sent-255, score-0.639]
98 At the centre of our idea is that the comparison of the treatment and control groups should be done in like-for-like fashion, giving rise to the partitioning of participants into sub-groups, each sub-group sharing the same intervention and post-trial responses. [sent-256, score-0.909]
99 A Bayesian framework was used to interpret jointly the auxiliary and trial outcome data, giving the clinician a meaningful and readily understandable end result. [sent-257, score-0.53]
100 Assessment of blinding may be inappropriate after the trial. [sent-293, score-0.593]
wordName wordTfidf (topN-words)
[('blinding', 0.593), ('participants', 0.398), ('treatment', 0.282), ('trial', 0.265), ('intervention', 0.159), ('bang', 0.142), ('blinded', 0.129), ('fig', 0.105), ('mcg', 0.103), ('unblinded', 0.103), ('subgroups', 0.091), ('unblinding', 0.09), ('group', 0.09), ('posteriors', 0.084), ('indexes', 0.084), ('questionnaire', 0.08), ('james', 0.079), ('dg', 0.079), ('pc', 0.078), ('decisive', 0.077), ('placebo', 0.077), ('auxiliary', 0.077), ('clinical', 0.076), ('participant', 0.075), ('assignment', 0.069), ('belief', 0.068), ('individuals', 0.065), ('administered', 0.064), ('xcg', 0.064), ('outcome', 0.064), ('effectiveness', 0.063), ('pt', 0.059), ('feedback', 0.058), ('clinician', 0.057), ('subgroup', 0.057), ('assigned', 0.056), ('assessment', 0.052), ('aforesaid', 0.052), ('blindness', 0.052), ('ncg', 0.052), ('control', 0.049), ('response', 0.047), ('index', 0.046), ('differential', 0.046), ('gt', 0.046), ('interventions', 0.046), ('responses', 0.044), ('regarding', 0.04), ('cg', 0.039), ('amalgamated', 0.039), ('clin', 0.039), ('contemp', 0.039), ('deems', 0.039), ('undecided', 0.039), ('progressive', 0.038), ('sec', 0.036), ('incorrect', 0.035), ('perfect', 0.034), ('hoc', 0.032), ('interpretation', 0.032), ('eq', 0.031), ('proportion', 0.03), ('lack', 0.029), ('insight', 0.029), ('et', 0.028), ('effect', 0.027), ('collected', 0.027), ('nature', 0.026), ('disclosed', 0.026), ('judgement', 0.026), ('mayberg', 0.026), ('pag', 0.026), ('signify', 0.026), ('xag', 0.026), ('posterior', 0.025), ('firstly', 0.025), ('meaningful', 0.025), ('weighting', 0.025), ('effects', 0.024), ('inherent', 0.024), ('nt', 0.024), ('conservative', 0.023), ('inherently', 0.023), ('atomization', 0.023), ('indecisive', 0.023), ('trials', 0.022), ('deemed', 0.022), ('assessing', 0.022), ('readily', 0.022), ('believes', 0.021), ('fat', 0.021), ('groups', 0.021), ('interpret', 0.02), ('quantify', 0.02), ('know', 0.02), ('dcg', 0.02), ('gc', 0.02), ('guesses', 0.02), ('solid', 0.02), ('extent', 0.019)]
simIndex simValue paperId paperTitle
same-paper 1 1.0000002 46 nips-2012-Assessing Blinding in Clinical Trials
Author: Ognjen Arandjelovic
Abstract: The interaction between the patient’s expected outcome of an intervention and the inherent effects of that intervention can have extraordinary effects. Thus in clinical trials an effort is made to conceal the nature of the administered intervention from the participants in the trial i.e. to blind it. Yet, in practice perfect blinding is impossible to ensure or even verify. The current standard is follow up the trial with an auxiliary questionnaire, which allows trial participants to express their belief concerning the assigned intervention and which is used to compute a measure of the extent of blinding in the trial. If the estimated extent of blinding exceeds a threshold the trial is deemed sufficiently blinded; otherwise, the trial is deemed to have failed. In this paper we make several important contributions. Firstly, we identify a series of fundamental problems of the aforesaid practice and discuss them in context of the most commonly used blinding measures. Secondly, motivated by the highlighted problems, we formulate a novel method for handling imperfectly blinded trials. We too adopt a post-trial feedback questionnaire but interpret the collected data using an original approach, fundamentally different from those previously proposed. Unlike previous approaches, ours is void of any ad hoc free parameters, is robust to small changes in auxiliary data and is not predicated on any strong assumptions used to interpret participants’ feedback. 1
2 0.054344155 273 nips-2012-Predicting Action Content On-Line and in Real Time before Action Onset – an Intracranial Human Study
Author: Uri Maoz, Shengxuan Ye, Ian Ross, Adam Mamelak, Christof Koch
Abstract: The ability to predict action content from neural signals in real time before the action occurs has been long sought in the neuroscientific study of decision-making, agency and volition. On-line real-time (ORT) prediction is important for understanding the relation between neural correlates of decision-making and conscious, voluntary action as well as for brain-machine interfaces. Here, epilepsy patients, implanted with intracranial depth microelectrodes or subdural grid electrodes for clinical purposes, participated in a “matching-pennies” game against an opponent. In each trial, subjects were given a 5 s countdown, after which they had to raise their left or right hand immediately as the “go” signal appeared on a computer screen. They won a fixed amount of money if they raised a different hand than their opponent and lost that amount otherwise. The question we here studied was the extent to which neural precursors of the subjects’ decisions can be detected in intracranial local field potentials (LFP) prior to the onset of the action. We found that combined low-frequency (0.1–5 Hz) LFP signals from 10 electrodes were predictive of the intended left-/right-hand movements before the onset of the go signal. Our ORT system predicted which hand the patient would raise 0.5 s before the go signal with 68±3% accuracy in two patients. Based on these results, we constructed an ORT system that tracked up to 30 electrodes simultaneously, and tested it on retrospective data from 7 patients. On average, we could predict the correct hand choice in 83% of the trials, which rose to 92% if we let the system drop 3/10 of the trials on which it was less confident. Our system demonstrates— for the first time—the feasibility of accurately predicting a binary action on single trials in real time for patients with intracranial recordings, well before the action occurs. 1 1
3 0.050613586 309 nips-2012-Semi-supervised Eigenvectors for Locally-biased Learning
Author: Toke Hansen, Michael W. Mahoney
Abstract: In many applications, one has side information, e.g., labels that are provided in a semi-supervised manner, about a specific target region of a large data set, and one wants to perform machine learning and data analysis tasks “nearby” that pre-specified target region. Locally-biased problems of this sort are particularly challenging for popular eigenvector-based machine learning and data analysis tools. At root, the reason is that eigenvectors are inherently global quantities. In this paper, we address this issue by providing a methodology to construct semi-supervised eigenvectors of a graph Laplacian, and we illustrate how these locally-biased eigenvectors can be used to perform locally-biased machine learning. These semi-supervised eigenvectors capture successively-orthogonalized directions of maximum variance, conditioned on being well-correlated with an input seed set of nodes that is assumed to be provided in a semi-supervised manner. We also provide several empirical examples demonstrating how these semi-supervised eigenvectors can be used to perform locally-biased learning. 1
4 0.049381156 64 nips-2012-Calibrated Elastic Regularization in Matrix Completion
Author: Tingni Sun, Cun-hui Zhang
Abstract: This paper concerns the problem of matrix completion, which is to estimate a matrix from observations in a small subset of indices. We propose a calibrated spectrum elastic net method with a sum of the nuclear and Frobenius penalties and develop an iterative algorithm to solve the convex minimization problem. The iterative algorithm alternates between imputing the missing entries in the incomplete matrix by the current guess and estimating the matrix by a scaled soft-thresholding singular value decomposition of the imputed matrix until the resulting matrix converges. A calibration step follows to correct the bias caused by the Frobenius penalty. Under proper coherence conditions and for suitable penalties levels, we prove that the proposed estimator achieves an error bound of nearly optimal order and in proportion to the noise level. This provides a unified analysis of the noisy and noiseless matrix completion problems. Simulation results are presented to compare our proposal with previous ones. 1
5 0.045873318 138 nips-2012-Fully Bayesian inference for neural models with negative-binomial spiking
Author: James Scott, Jonathan W. Pillow
Abstract: Characterizing the information carried by neural populations in the brain requires accurate statistical models of neural spike responses. The negative-binomial distribution provides a convenient model for over-dispersed spike counts, that is, responses with greater-than-Poisson variability. Here we describe a powerful data-augmentation framework for fully Bayesian inference in neural models with negative-binomial spiking. Our approach relies on a recently described latentvariable representation of the negative-binomial distribution, which equates it to a Polya-gamma mixture of normals. This framework provides a tractable, conditionally Gaussian representation of the posterior that can be used to design efficient EM and Gibbs sampling based algorithms for inference in regression and dynamic factor models. We apply the model to neural data from primate retina and show that it substantially outperforms Poisson regression on held-out data, and reveals latent structure underlying spike count correlations in simultaneously recorded spike trains. 1
7 0.042591538 56 nips-2012-Bayesian active learning with localized priors for fast receptive field characterization
8 0.042509876 238 nips-2012-Neurally Plausible Reinforcement Learning of Working Memory Tasks
9 0.040571701 153 nips-2012-How Prior Probability Influences Decision Making: A Unifying Probabilistic Model
10 0.040351439 348 nips-2012-Tractable Objectives for Robust Policy Optimization
11 0.038980957 155 nips-2012-Human memory search as a random walk in a semantic network
12 0.036336739 83 nips-2012-Controlled Recognition Bounds for Visual Learning and Exploration
13 0.035627995 13 nips-2012-A Nonparametric Conjugate Prior Distribution for the Maximizing Argument of a Noisy Function
14 0.033828955 104 nips-2012-Dual-Space Analysis of the Sparse Linear Model
15 0.032172896 60 nips-2012-Bayesian nonparametric models for ranked data
16 0.03123853 77 nips-2012-Complex Inference in Neural Circuits with Probabilistic Population Codes and Topic Models
17 0.030947894 111 nips-2012-Efficient Sampling for Bipartite Matching Problems
18 0.029332073 16 nips-2012-A Polynomial-time Form of Robust Regression
19 0.02909359 252 nips-2012-On Multilabel Classification and Ranking with Partial Feedback
20 0.028014265 121 nips-2012-Expectation Propagation in Gaussian Process Dynamical Systems
topicId topicWeight
[(0, 0.091), (1, -0.002), (2, -0.01), (3, 0.041), (4, -0.019), (5, 0.007), (6, -0.005), (7, 0.013), (8, -0.007), (9, 0.019), (10, -0.014), (11, -0.011), (12, -0.027), (13, -0.012), (14, -0.002), (15, 0.002), (16, -0.006), (17, -0.007), (18, -0.007), (19, -0.019), (20, -0.006), (21, 0.025), (22, -0.019), (23, -0.024), (24, 0.029), (25, -0.043), (26, 0.001), (27, 0.053), (28, -0.003), (29, 0.02), (30, 0.01), (31, -0.023), (32, -0.037), (33, -0.022), (34, 0.036), (35, -0.006), (36, 0.007), (37, -0.036), (38, -0.052), (39, -0.014), (40, -0.013), (41, -0.002), (42, 0.023), (43, 0.037), (44, -0.011), (45, -0.011), (46, 0.002), (47, 0.037), (48, -0.035), (49, -0.056)]
simIndex simValue paperId paperTitle
same-paper 1 0.89466923 46 nips-2012-Assessing Blinding in Clinical Trials
Author: Ognjen Arandjelovic
Abstract: The interaction between the patient’s expected outcome of an intervention and the inherent effects of that intervention can have extraordinary effects. Thus in clinical trials an effort is made to conceal the nature of the administered intervention from the participants in the trial i.e. to blind it. Yet, in practice perfect blinding is impossible to ensure or even verify. The current standard is follow up the trial with an auxiliary questionnaire, which allows trial participants to express their belief concerning the assigned intervention and which is used to compute a measure of the extent of blinding in the trial. If the estimated extent of blinding exceeds a threshold the trial is deemed sufficiently blinded; otherwise, the trial is deemed to have failed. In this paper we make several important contributions. Firstly, we identify a series of fundamental problems of the aforesaid practice and discuss them in context of the most commonly used blinding measures. Secondly, motivated by the highlighted problems, we formulate a novel method for handling imperfectly blinded trials. We too adopt a post-trial feedback questionnaire but interpret the collected data using an original approach, fundamentally different from those previously proposed. Unlike previous approaches, ours is void of any ad hoc free parameters, is robust to small changes in auxiliary data and is not predicated on any strong assumptions used to interpret participants’ feedback. 1
Author: Won H. Kim, Deepti Pachauri, Charles Hatt, Moo. K. Chung, Sterling Johnson, Vikas Singh
Abstract: Hypothesis testing on signals defined on surfaces (such as the cortical surface) is a fundamental component of a variety of studies in Neuroscience. The goal here is to identify regions that exhibit changes as a function of the clinical condition under study. As the clinical questions of interest move towards identifying very early signs of diseases, the corresponding statistical differences at the group level invariably become weaker and increasingly hard to identify. Indeed, after a multiple comparisons correction is adopted (to account for correlated statistical tests over all surface points), very few regions may survive. In contrast to hypothesis tests on point-wise measurements, in this paper, we make the case for performing statistical analysis on multi-scale shape descriptors that characterize the local topological context of the signal around each surface vertex. Our descriptors are based on recent results from harmonic analysis, that show how wavelet theory extends to non-Euclidean settings (i.e., irregular weighted graphs). We provide strong evidence that these descriptors successfully pick up group-wise differences, where traditional methods either fail or yield unsatisfactory results. Other than this primary application, we show how the framework allows performing cortical surface smoothing in the native space without mappint to a unit sphere. 1
3 0.59227431 276 nips-2012-Probabilistic Event Cascades for Alzheimer's disease
Author: Jonathan Huang, Daniel Alexander
Abstract: Accurate and detailed models of neurodegenerative disease progression are crucially important for reliable early diagnosis and the determination of effective treatments. We introduce the ALPACA (Alzheimer’s disease Probabilistic Cascades) model, a generative model linking latent Alzheimer’s progression dynamics to observable biomarker data. In contrast with previous works which model disease progression as a fixed event ordering, we explicitly model the variability over such orderings among patients which is more realistic, particularly for highly detailed progression models. We describe efficient learning algorithms for ALPACA and discuss promising experimental results on a real cohort of Alzheimer’s patients from the Alzheimer’s Disease Neuroimaging Initiative. 1
Author: Uri Maoz, Shengxuan Ye, Ian Ross, Adam Mamelak, Christof Koch
Abstract: The ability to predict action content from neural signals in real time before the action occurs has been long sought in the neuroscientific study of decision-making, agency and volition. On-line real-time (ORT) prediction is important for understanding the relation between neural correlates of decision-making and conscious, voluntary action as well as for brain-machine interfaces. Here, epilepsy patients, implanted with intracranial depth microelectrodes or subdural grid electrodes for clinical purposes, participated in a “matching-pennies” game against an opponent. In each trial, subjects were given a 5 s countdown, after which they had to raise their left or right hand immediately as the “go” signal appeared on a computer screen. They won a fixed amount of money if they raised a different hand than their opponent and lost that amount otherwise. The question we here studied was the extent to which neural precursors of the subjects’ decisions can be detected in intracranial local field potentials (LFP) prior to the onset of the action. We found that combined low-frequency (0.1–5 Hz) LFP signals from 10 electrodes were predictive of the intended left-/right-hand movements before the onset of the go signal. Our ORT system predicted which hand the patient would raise 0.5 s before the go signal with 68±3% accuracy in two patients. Based on these results, we constructed an ORT system that tracked up to 30 electrodes simultaneously, and tested it on retrospective data from 7 patients. On average, we could predict the correct hand choice in 83% of the trials, which rose to 92% if we let the system drop 3/10 of the trials on which it was less confident. Our system demonstrates— for the first time—the feasibility of accurately predicting a binary action on single trials in real time for patients with intracranial recordings, well before the action occurs. 1 1
5 0.54288793 155 nips-2012-Human memory search as a random walk in a semantic network
Author: Joseph L. Austerweil, Joshua T. Abbott, Thomas L. Griffiths
Abstract: The human mind has a remarkable ability to store a vast amount of information in memory, and an even more remarkable ability to retrieve these experiences when needed. Understanding the representations and algorithms that underlie human memory search could potentially be useful in other information retrieval settings, including internet search. Psychological studies have revealed clear regularities in how people search their memory, with clusters of semantically related items tending to be retrieved together. These findings have recently been taken as evidence that human memory search is similar to animals foraging for food in patchy environments, with people making a rational decision to switch away from a cluster of related information as it becomes depleted. We demonstrate that the results that were taken as evidence for this account also emerge from a random walk on a semantic network, much like the random web surfer model used in internet search engines. This offers a simpler and more unified account of how people search their memory, postulating a single process rather than one process for exploring a cluster and one process for switching between clusters. 1
7 0.52663606 53 nips-2012-Bayesian Pedigree Analysis using Measure Factorization
9 0.51822531 157 nips-2012-Identification of Recurrent Patterns in the Activation of Brain Networks
10 0.4813396 94 nips-2012-Delay Compensation with Dynamical Synapses
11 0.47270906 198 nips-2012-Learning with Target Prior
12 0.45747349 238 nips-2012-Neurally Plausible Reinforcement Learning of Working Memory Tasks
13 0.45462298 256 nips-2012-On the connections between saliency and tracking
14 0.44295964 212 nips-2012-Minimax Multi-Task Learning and a Generalized Loss-Compositional Paradigm for MTL
15 0.44270328 28 nips-2012-A systematic approach to extracting semantic information from functional MRI data
16 0.43143895 23 nips-2012-A lattice filter model of the visual pathway
17 0.42967767 24 nips-2012-A mechanistic model of early sensory processing based on subtracting sparse representations
18 0.41696143 333 nips-2012-Synchronization can Control Regularization in Neural Systems via Correlated Noise Processes
19 0.40384072 77 nips-2012-Complex Inference in Neural Circuits with Probabilistic Population Codes and Topic Models
20 0.40022388 137 nips-2012-From Deformations to Parts: Motion-based Segmentation of 3D Objects
topicId topicWeight
[(0, 0.046), (11, 0.015), (17, 0.016), (21, 0.032), (38, 0.096), (42, 0.028), (54, 0.032), (55, 0.036), (74, 0.035), (76, 0.134), (80, 0.069), (91, 0.331), (92, 0.033)]
simIndex simValue paperId paperTitle
same-paper 1 0.76963907 46 nips-2012-Assessing Blinding in Clinical Trials
Author: Ognjen Arandjelovic
Abstract: The interaction between the patient’s expected outcome of an intervention and the inherent effects of that intervention can have extraordinary effects. Thus in clinical trials an effort is made to conceal the nature of the administered intervention from the participants in the trial i.e. to blind it. Yet, in practice perfect blinding is impossible to ensure or even verify. The current standard is follow up the trial with an auxiliary questionnaire, which allows trial participants to express their belief concerning the assigned intervention and which is used to compute a measure of the extent of blinding in the trial. If the estimated extent of blinding exceeds a threshold the trial is deemed sufficiently blinded; otherwise, the trial is deemed to have failed. In this paper we make several important contributions. Firstly, we identify a series of fundamental problems of the aforesaid practice and discuss them in context of the most commonly used blinding measures. Secondly, motivated by the highlighted problems, we formulate a novel method for handling imperfectly blinded trials. We too adopt a post-trial feedback questionnaire but interpret the collected data using an original approach, fundamentally different from those previously proposed. Unlike previous approaches, ours is void of any ad hoc free parameters, is robust to small changes in auxiliary data and is not predicated on any strong assumptions used to interpret participants’ feedback. 1
2 0.69265753 207 nips-2012-Mandatory Leaf Node Prediction in Hierarchical Multilabel Classification
Author: Wei Bi, James T. Kwok
Abstract: In hierarchical classification, the prediction paths may be required to always end at leaf nodes. This is called mandatory leaf node prediction (MLNP) and is particularly useful when the leaf nodes have much stronger semantic meaning than the internal nodes. However, while there have been a lot of MLNP methods in hierarchical multiclass classification, performing MLNP in hierarchical multilabel classification is much more difficult. In this paper, we propose a novel MLNP algorithm that (i) considers the global hierarchy structure; and (ii) can be used on hierarchies of both trees and DAGs. We show that one can efficiently maximize the joint posterior probability of all the node labels by a simple greedy algorithm. Moreover, this can be further extended to the minimization of the expected symmetric loss. Experiments are performed on a number of real-world data sets with tree- and DAG-structured label hierarchies. The proposed method consistently outperforms other hierarchical and flat multilabel classification methods. 1
3 0.60689998 122 nips-2012-Exploration in Model-based Reinforcement Learning by Empirically Estimating Learning Progress
Author: Manuel Lopes, Tobias Lang, Marc Toussaint, Pierre-yves Oudeyer
Abstract: Formal exploration approaches in model-based reinforcement learning estimate the accuracy of the currently learned model without consideration of the empirical prediction error. For example, PAC-MDP approaches such as R- MAX base their model certainty on the amount of collected data, while Bayesian approaches assume a prior over the transition dynamics. We propose extensions to such approaches which drive exploration solely based on empirical estimates of the learner’s accuracy and learning progress. We provide a “sanity check” theoretical analysis, discussing the behavior of our extensions in the standard stationary finite state-action case. We then provide experimental studies demonstrating the robustness of these exploration measures in cases of non-stationary environments or where original approaches are misled by wrong domain assumptions. 1
4 0.50465035 354 nips-2012-Truly Nonparametric Online Variational Inference for Hierarchical Dirichlet Processes
Author: Michael Bryant, Erik B. Sudderth
Abstract: Variational methods provide a computationally scalable alternative to Monte Carlo methods for large-scale, Bayesian nonparametric learning. In practice, however, conventional batch and online variational methods quickly become trapped in local optima. In this paper, we consider a nonparametric topic model based on the hierarchical Dirichlet process (HDP), and develop a novel online variational inference algorithm based on split-merge topic updates. We derive a simpler and faster variational approximation of the HDP, and show that by intelligently splitting and merging components of the variational posterior, we can achieve substantially better predictions of test data than conventional online and batch variational algorithms. For streaming analysis of large datasets where batch analysis is infeasible, we show that our split-merge updates better capture the nonparametric properties of the underlying model, allowing continual learning of new topics.
5 0.50290889 104 nips-2012-Dual-Space Analysis of the Sparse Linear Model
Author: Yi Wu, David P. Wipf
Abstract: Sparse linear (or generalized linear) models combine a standard likelihood function with a sparse prior on the unknown coefficients. These priors can conveniently be expressed as a maximization over zero-mean Gaussians with different variance hyperparameters. Standard MAP estimation (Type I) involves maximizing over both the hyperparameters and coefficients, while an empirical Bayesian alternative (Type II) first marginalizes the coefficients and then maximizes over the hyperparameters, leading to a tractable posterior approximation. The underlying cost functions can be related via a dual-space framework from [22], which allows both the Type I or Type II objectives to be expressed in either coefficient or hyperparmeter space. This perspective is useful because some analyses or extensions are more conducive to development in one space or the other. Herein we consider the estimation of a trade-off parameter balancing sparsity and data fit. As this parameter is effectively a variance, natural estimators exist by assessing the problem in hyperparameter (variance) space, transitioning natural ideas from Type II to solve what is much less intuitive for Type I. In contrast, for analyses of update rules and sparsity properties of local and global solutions, as well as extensions to more general likelihood models, we can leverage coefficient-space techniques developed for Type I and apply them to Type II. For example, this allows us to prove that Type II-inspired techniques can be successful recovering sparse coefficients when unfavorable restricted isometry properties (RIP) lead to failure of popular ℓ1 reconstructions. It also facilitates the analysis of Type II when non-Gaussian likelihood models lead to intractable integrations. 1
6 0.5018605 188 nips-2012-Learning from Distributions via Support Measure Machines
7 0.50171781 316 nips-2012-Small-Variance Asymptotics for Exponential Family Dirichlet Process Mixture Models
8 0.4988713 209 nips-2012-Max-Margin Structured Output Regression for Spatio-Temporal Action Localization
9 0.49832013 83 nips-2012-Controlled Recognition Bounds for Visual Learning and Exploration
10 0.49830672 318 nips-2012-Sparse Approximate Manifolds for Differential Geometric MCMC
11 0.49818495 294 nips-2012-Repulsive Mixtures
12 0.49816275 232 nips-2012-Multiplicative Forests for Continuous-Time Processes
13 0.49804756 355 nips-2012-Truncation-free Online Variational Inference for Bayesian Nonparametric Models
14 0.49790505 215 nips-2012-Minimizing Uncertainty in Pipelines
15 0.49763227 284 nips-2012-Q-MKL: Matrix-induced Regularization in Multi-Kernel Learning with Applications to Neuroimaging
16 0.49761656 272 nips-2012-Practical Bayesian Optimization of Machine Learning Algorithms
17 0.49755719 298 nips-2012-Scalable Inference of Overlapping Communities
18 0.49746996 353 nips-2012-Transferring Expectations in Model-based Reinforcement Learning
19 0.49661231 68 nips-2012-Clustering Aggregation as Maximum-Weight Independent Set
20 0.49647674 126 nips-2012-FastEx: Hash Clustering with Exponential Families