-
Conditional Sequential Slate Optimization
Authors:
Yipeng Zhang,
Mingjian Lu,
Saratchandra Indrakanti,
Manojkumar Rangasamy Kannadasan,
Abraham Bagherjeiran
Abstract:
The top search results matching a user query that are displayed on the first page are critical to the effectiveness and perception of a search system. A search ranking system typically orders the results by independent query-document scores to produce a slate of search results. However, such unilateral scoring methods may fail to capture inter-document dependencies that users are sensitive to, thu…
▽ More
The top search results matching a user query that are displayed on the first page are critical to the effectiveness and perception of a search system. A search ranking system typically orders the results by independent query-document scores to produce a slate of search results. However, such unilateral scoring methods may fail to capture inter-document dependencies that users are sensitive to, thus producing a sub-optimal slate. Further, in practice, many real-world applications such as e-commerce search require enforcing certain distributional criteria at the slate-level, due to business objectives or long term user retention goals. Unilateral scoring of results does not explicitly support optimizing for such objectives with respect to a slate. Hence, solutions to the slate optimization problem must consider the optimal selection and order of the documents, along with adherence to slate-level distributional criteria. To that end, we propose a hybrid framework extended from traditional slate optimization to solve the conditional slate optimization problem. We introduce conditional sequential slate optimization (CSSO), which jointly learns to optimize for traditional ranking metrics as well as prescribed distribution criteria of documents within the slate. The proposed method can be applied to practical real world problems such as enforcing diversity in e-commerce search results, mitigating bias in top results and personalization of results. Experiments on public datasets and real-world data from e-commerce datasets show that CSSO outperforms popular comparable ranking methods in terms of adherence to distributional criteria while producing comparable or better relevance metrics.
△ Less
Submitted 13 August, 2021; v1 submitted 12 August, 2021;
originally announced August 2021.
-
Addressing Purchase-Impression Gap through a Sequential Re-ranker
Authors:
Shubhangi Tandon,
Saratchandra Indrakanti,
Amit Jaiswal,
Svetlana Strunjas,
Manojkumar Rangasamy Kannadasan
Abstract:
Large scale eCommerce platforms such as eBay carry a wide variety of inventory and provide several buying choices to online shoppers. It is critical for eCommerce search engines to showcase in the top results the variety and selection of inventory available, specifically in the context of the various buying intents that may be associated with a search query. Search rankers are most commonly powere…
▽ More
Large scale eCommerce platforms such as eBay carry a wide variety of inventory and provide several buying choices to online shoppers. It is critical for eCommerce search engines to showcase in the top results the variety and selection of inventory available, specifically in the context of the various buying intents that may be associated with a search query. Search rankers are most commonly powered by learning-to-rank models which learn the preference between items during training. However, they score items independent of other items at runtime. Although the items placed at top of the results by such scoring functions may be independently optimal, they can be sub-optimal as a set. This may lead to a mismatch between the ideal distribution of items in the top results vs what is actually impressed. In this paper, we present methods to address the purchase-impression gap observed in top search results on eCommerce sites. We establish the ideal distribution of items based on historic shopping patterns. We then present a sequential reranker that methodically reranks top search results produced by a conventional pointwise scoring ranker. The reranker produces a reordered list by sequentially selecting candidates trading off between their independent relevance and potential to address the purchase-impression gap by utilizing specially constructed features that capture impression distribution of items already added to a reranked list. The sequential reranker enables addressing purchase impression gap with respect to multiple item aspects. Early version of the reranker showed promising lifts in conversion and engagement metrics at eBay. Based on experiments on randomly sampled validation datasets, we observe that the reranking methodology presented produces around 10% reduction in purchase-impression gap at an average for the top 20 results, while making improvements to conversion metrics.
△ Less
Submitted 27 October, 2020;
originally announced October 2020.
-
Influence of Neighborhood on the Preference of an Item in eCommerce Search
Authors:
Saratchandra Indrakanti,
Svetlana Strunjas,
Shubhangi Tandon,
Manojkumar Rangasamy Kannadasan
Abstract:
Surfacing a ranked list of items for a search query to help buyers discover inventory and make purchase decisions is a critical problem in eCommerce search. Typically, items are independently predicted with a probability of sale with respect to a given search query. But in a dynamic marketplace like eBay, even for a single product, there are various different factors distinguishing one item from a…
▽ More
Surfacing a ranked list of items for a search query to help buyers discover inventory and make purchase decisions is a critical problem in eCommerce search. Typically, items are independently predicted with a probability of sale with respect to a given search query. But in a dynamic marketplace like eBay, even for a single product, there are various different factors distinguishing one item from another which can influence the purchase decision for the user. Users have to make a purchase decision by considering all of these options. Majority of the existing learning to rank algorithms model the relative relevance between labeled items only at the loss functions like pairwise or list-wise losses. But they are limited to point-wise scoring functions where items are ranked independently based on the features of the item itself. In this paper, we study the influence of an item's neighborhood to its purchase decision. Here, we consider the neighborhood as the items ranked above and below the current item in search results. By adding delta features comparing items within a neighborhood and learning a ranking model, we are able to experimentally show that the new ranker with delta features outperforms our baseline ranker in terms of Mean Reciprocal Rank (MRR). The ranking models with proposed delta features result in $3-5\%$ improvement in MRR over the baseline model. We also study impact of different sizes for neighborhood. Experimental results show that neighborhood size $3$ perform the best based on MRR with an improvement of $4-5\%$ over the baseline model.
△ Less
Submitted 17 October, 2019; v1 submitted 10 August, 2019;
originally announced August 2019.
-
Large field generation with Hot Isostatically Pressed Powder-in-Tube MgB2 coil at 25 K
Authors:
A. Serquis,
L. Civale,
J. Y. Coulter,
D. L. Hammon,
X. Z. Liao,
Y. T. Zhu,
D. E. Peterson,
F. M. Mueller,
V. F. Nesterenko,
S. S. Indrakanti
Abstract:
We present the fabrication and test results of Hot-Isostatic-Pressed (HIPed) Powder-in-Tube (PIT) MgB$_2$ coils. The coils properties were measured by transport and magnetization at different applied fields ($H$) and temperatures ($T$). The engineering critical current ($J_e$) value is the largest reported in PIT MgB$_2$ wires or tapes. At 25 K our champion 6-layer coil was able to generate a fi…
▽ More
We present the fabrication and test results of Hot-Isostatic-Pressed (HIPed) Powder-in-Tube (PIT) MgB$_2$ coils. The coils properties were measured by transport and magnetization at different applied fields ($H$) and temperatures ($T$). The engineering critical current ($J_e$) value is the largest reported in PIT MgB$_2$ wires or tapes. At 25 K our champion 6-layer coil was able to generate a field of 1 T at self-field ($I_c >$ 220 A, $J_e \sim 2.8 \times 10^4$ A/cm$^2$). At 4 K this coil generated 1.6 T under an applied field of 1.25 T ($I_c \sim350$ A, $J_e \sim 4.5 \times 10^4$ A/cm$^2$). These magnetic fields are high enough for a superconducting transformer or magnet applications such as MRI. A SiC doped MgB$_2$ single layer coil shows a promising improvement at high fields and exhibits $J_c > 10^4$ A/cm$^2$ at 7 T.
△ Less
Submitted 23 April, 2004; v1 submitted 2 April, 2004;
originally announced April 2004.
-
The degradation of MgB2 under ambient environment
Authors:
A. Serquis,
Y. T. Zhu,
D. E. Peterson,
F. M. Mueller,
R. K. Schulze,
V. F. Nesterenko,
S. S. Indrakanti
Abstract:
The superconductivities of samples prepared by several procedures were found to degrade under ambient environment. The degradation mechanism was studied by measuring the change of surface chemical composition of dense MgB2 pellets (prepared by hot isostatic pressure, HIPed) under atmospheric exposure using X-ray Photoelectron Spectroscopy (XPS). Results showed that samples with poor connectivity…
▽ More
The superconductivities of samples prepared by several procedures were found to degrade under ambient environment. The degradation mechanism was studied by measuring the change of surface chemical composition of dense MgB2 pellets (prepared by hot isostatic pressure, HIPed) under atmospheric exposure using X-ray Photoelectron Spectroscopy (XPS). Results showed that samples with poor connectivity between grains and with smaller grain sizes degrade with time when exposed to ambient conditions. In these samples, the Tc did not change with time, but the superconducting transition became broader and the Meissner fraction decreased. In contrast, our well-sintered and the HIPed samples remained stable for several months under ambient condition. The degradation was found to be related to surface decomposition as observed by XPS. We observed the formation of oxidized Mg, primarily in the form of a Mg hydroxide, the increase of C and O contents, and the reduction of B concentration in the surface layer of MgB2 samples.
△ Less
Submitted 25 June, 2002;
originally announced June 2002.
-
The influence of microstructures and crystalline defects on the superconductivity of MgB2
Authors:
A. Serquis,
X. Y. Liao,
Y. T. Zhu,
J. Y. Coulter,
J. Y. Huang,
J. O. Willis,
D. E. Peterson,
F. M. Mueller,
N. O. Moreno,
J. D. Thompson,
S. S. Indrakanti,
V. F. Nesterenko
Abstract:
This work studies the influence of microstructures and crystalline defects on the superconductivity of MgB2, with the objective to improve its flux pinning. A MgB2 sample pellet that was hot isostatic pressed (HIPed) was found to have significantly increased critical current density (Jc) at high fields than its un-HIPed counterpart. The HIPed sample had a Jc of 10000 A/cm2 in 50000 Oe (5 T) at 5…
▽ More
This work studies the influence of microstructures and crystalline defects on the superconductivity of MgB2, with the objective to improve its flux pinning. A MgB2 sample pellet that was hot isostatic pressed (HIPed) was found to have significantly increased critical current density (Jc) at high fields than its un-HIPed counterpart. The HIPed sample had a Jc of 10000 A/cm2 in 50000 Oe (5 T) at 5K. This was 20 times higher than that of the un-HIPed sample, and the same as the best Jc reported by other research groups. Microstructures observed in scanning and transmission electron microscopy indicate that the HIP process eliminated porosity present in the MgB2 pellet resulting in an improved intergrain connectivity. Such improvement in intergrain connectivity was believed to prevent the steep Jc drop with magnetic field H that occurred in the un-HIPed MgB2 pellet at H > 45000 Oe (4.5 T) and T = 5 K. The HIP process was also found to disperse the MgO that existed at the grain boundaries of the un-HIPed MgB2 pellet and to generate more dislocations in the HIPed the pellets. These dispersed MgO particles and dislocations improved flux pinning also at H<45000 Oe. The HIPing process was also found to lower the resistivity at room temperature.
△ Less
Submitted 5 March, 2002; v1 submitted 25 January, 2002;
originally announced January 2002.
-
Synthesis, characterization and ageing of MgB2
Authors:
A. Serquis,
R. Schulze,
Y. T. Zhu,
J. Y. Coulter,
D. E. Peterson,
N. O. Moreno,
P. G. Pagliuso,
S. S. Indrakanti,
V. F. Nesterenko,
F. M. Mueller
Abstract:
We studied the influence of sample preparation and defects in the superconducting properties samples using atomic ratios of Mg:B=1:1 and Mg:B=1:2. Samples were characterized by SEM, and XRD, and the magnetization properties were examined in a SQUID magnetometer. The presence of Mg vacancies was determined by Rietveld analysis. Most of the samples exhibited sharp superconducting transitions with…
▽ More
We studied the influence of sample preparation and defects in the superconducting properties samples using atomic ratios of Mg:B=1:1 and Mg:B=1:2. Samples were characterized by SEM, and XRD, and the magnetization properties were examined in a SQUID magnetometer. The presence of Mg vacancies was determined by Rietveld analysis. Most of the samples exhibited sharp superconducting transitions with Tcs between 37- 39 K. We found a strong correlation between the crystal strain and the Tc. This strain was related to the presence of Mg vacancies. In addition, results showed that some samples degraded with time when exposed to ambient conditions. In these samples the Tc did not change with time, but the superconducting transition became broader and the Meissner fraction decreased. This effect was only present in samples with poor connectivity between grains and smaller grain sizes. The degradation was related to a surface decomposition as observed by X-ray Photoelectron Spectroscopy. No correlation was found between this effect and the presence of Mg vacancies.
△ Less
Submitted 13 December, 2001;
originally announced December 2001.
-
Improved superconducting properties of MgB2
Authors:
N. A. Frederick,
Shi Li,
M. B. Maple,
V. F. Nesterenko,
S. S. Indrakanti
Abstract:
We present electrical transport, magnetization, and specific heat measurements on bulk MgB2 samples (T_{c} = 38.5 K) synthesized under 200 MPa pressure using a process based on hot isostatic pressing with cooling under pressure. The samples are fully dense and display excellent superconducting properties, including a narrow superconducting transition width (ΔT_{c} = 0.75 K), a high upper critica…
▽ More
We present electrical transport, magnetization, and specific heat measurements on bulk MgB2 samples (T_{c} = 38.5 K) synthesized under 200 MPa pressure using a process based on hot isostatic pressing with cooling under pressure. The samples are fully dense and display excellent superconducting properties, including a narrow superconducting transition width (ΔT_{c} = 0.75 K), a high upper critical field H_{c2} (H_{c2}(0) ~ 155 kOe), and a critical current density J_{c} that is the largest yet measured for bulk samples of MgB2 (J_{c}(0) ~ 1.4 MA/cm^{2}). Specific heat measurements yielded a jump ΔC at T_{c} of 92 mJ/mol K. These superconducting properties are comparable to those obtained with techniques that are not so well suited to industrial scale fabrication.
△ Less
Submitted 25 June, 2001;
originally announced June 2001.
-
Hot isostatic pressing of bulk magnesium diboride: mechanical and superconducting properties
Authors:
S. S. Indrakanti,
V. F. Nesterenko,
M. B. Maple,
N. A. Frederick,
W. M. Yuhasz,
Shi Li
Abstract:
Two different hot isostatic pressing cycles (HIPing) were investigated to synthesize bulk MgB2 samples: a "standard" cycle where a low vessel pressure is maintained while heating to the process temperature with a subsequent simultaneous pressure and temperature decrease and a new method - dense material cooling under pressure (DMCUP). The latter method allowed the synthesis of dense samples with…
▽ More
Two different hot isostatic pressing cycles (HIPing) were investigated to synthesize bulk MgB2 samples: a "standard" cycle where a low vessel pressure is maintained while heating to the process temperature with a subsequent simultaneous pressure and temperature decrease and a new method - dense material cooling under pressure (DMCUP). The latter method allowed the synthesis of dense samples with diameters up to 20 mm and thicknesses up to 10 mm from commercial MgB2 powder. Optimal conditions for the DMCUP method with glass encapsulation (maximum pressure 200 MPa, maximum temperature 1000 C over 200 min, and cooling under pressure) resulted in a dense material with a sharp superconducting transition at 38.5 K. This method employs a pressure which is one order of magnitude less than previously reported for pressure assisted sintering of dense material and can be scaled to larger sample sizes and complex shapes. The data for density, microhardness, fracture toughness and sound speed as well as superconducting properties for bulk magnesium diboride are presented. Ball milling the powder enhances sintering and results in a more homogeneous final microstructure.
△ Less
Submitted 8 August, 2001; v1 submitted 24 May, 2001;
originally announced May 2001.
